In part one of this series, we looked at basic genetics concepts such as DNA, inheritance, and cells and cell division, and in part two we examined more difficult ideas, such as dominance and recessivity, sex-linked traits, genetic pathways, development and environmental effects. From here on, I'm going to assume you're comfortable with all of this, so please review if necessary.

So far, we've just considered situations where everything goes to plan, focussing mostly on theoretical concepts. But real life is messy, so let's conclude this series by switching our attention to some practical applications of genetics, by taking a brief look at chromosome variations, genetic mutation, cancer, selective breeding and disease genetics.

By this point, you should be able to visualize what's going on inside our cells at the genetic level. You know from part one that our 30 trillion body cells each carry a genome consisting of a DNA-protein complex, divided into 23 pairs of chromosomes: 22 autosome pairs plus a pair of sex chromosomes. You also know that our sex cells (eggs or sperm) each carry only one chromosome from each pair, with each single chromosome again becoming one member of a pair following fertilization (the fusing of the sperm with the egg cell to make a zygote).

Remember also from part one that, immediately prior to their separation during meiosis, the members of the chromosome pair exchange segments with one another, a mechanism which greatly increases genetic variation amongst offspring. This exchanging of segments is not quite perfect; there are several ways in which things can go wrong, either with the exchanging process or with the subsequent separation.

As we have seen, every chromosome pair will normally separate during metaphase, with one member of each pair moving into one daughter cell, and the other member of the pair moving into the other daughter cell. However, sometimes this separation does not work correctly for a pair of chromosomes, with the result that one daughter cell receives both members of the pair, while the other receives neither.

Left: normal meiosis for a single chromosome pair. Top, each chromosome in the pair has duplicated into connected sister chromatids and lined up ready for cell division at metaphase. Middle, the pair separates into separate daughter cells during meiosis I. Bottom, the sister chromatids separate during meiosis II. Right: non-disjunction meiosis. Normal metaphase (top) and meiosis I (middle), but non-disjunction occurs during meiosis II (bottom). This results in a gamete with a missing chromosome, and another gamete with an extra chromosome. Note that the other chromosomes are omitted for clarity, and that non-disjunction can also occur during meiosis I.

This leads to gametes with an extra or missing chromosome. When a gamete with two copies of a chromosome fuses with a normal gamete with one copy of the chromosome, at fertilization, the resulting zygote will have three copies of that chromosome (a trisomy). When a gamete with zero copies of a chromosome fuses with a normal gamete, the zygote will have only one copy of that chromosome (a monosomy). Although trisomies and monosomies can occur for any chromosome, the only ones compatible with life¹ are trisomy 21 (Down Syndrome) and those affecting the sex chromosomes.

An individual with Down Syndrome has three copies of chromosome 21, and two copies of the other chromosomes. This extra genetic material leads to the collection of attributes that are characteristic of the syndrome.

Trisomy 21 results from an extra copy of chromosome 21, as seen in the karyogram of this female.

Monosomies of the autosomes involve missing genes; all are too severe for a pregnancy to continue for more than a brief period.² Monosomy of the X-chromosome (XO) is called Turner Syndrome. There is one X-chromosome and no Y-chromosome. The effects of this are usually relatively mild, resulting in females with short stature and infertility but normal intelligence. Klinefelter syndrome (XXY) is another relatively mild condition, resulting in males with an extra X-chromosome.

Turner syndrome females have a single X-chromosome and no other sex chromosomes.

In addition to non-disjunction (leading to monosomies and trisomies), errors can also occur during crossing over (the exchange of materials between chromosomes that takes places prior to the chromosomes' separation, discussed in part two). The mechanism of crossing over, meiotic recombination, is highly complex and has long been the focus of intense research, but at a conceptual level it's easy to understand as the exchange of corresponding segments of DNA from two homologous chromosomes. When this process works correctly, it results in a shuffling of genes between two chromosomes. However, if the exchange is unequal, this can lead to duplication and/or deletion of a section of a chromosome.

Left: duplication occurs when a segment of DNA is repeated, giving extra copies of the genes contained within the duplicated segment. Right: deletion occurs when a segment of DNA is removed from a chromosome, losing the genes contained within the deleted segment.

Duplications and deletions can vary greatly in size, incorporating few or many genes. Clearly, the larger the duplication or deletion, the more genes will be affected and the greater the effect on the organism carrying it. Also, as we have seen, in general duplication is less damaging than deletion.

Since crossing over relies on temporary breakage of chromosomes, there is also a risk that when the chromosome segments are rejoined, they can rejoin the wrong way around. This is called a chromosomal inversion, and although it usually doesn't affect the carrier of the inversion, it does lead to a reduction in fertility in proportion to the size of the inversion (when crossing over occurs within the inverted segment, the resulting chromosomes are unbalanced).

A more complex situation occurs when crossing over mistakenly occurs between two non-homologous chromosomes, leading to an exchange of DNA sequence between different chromosomes. Many of these translocations occur sporadically, but some specific ones recur in certain diseases, including many cancers (more on which later).

Left: two different (non-homologous) chromosomes erroneously line up prior to crossing over. Center: a segment breaks off one member of each pair. Right: the non-homologous segments attach to different chromosomes. The resulting products are translocations.

Carriers of translocations usually have a balanced gene complement, but problems can occur when having children, the most obvious of which is reduced fertility. This reduction in fertility is due to the production of unbalanced gametes (i.e. gametes that have an extra segment of one chromosome and a deleted segment of another chromosome), which usually produce an unviable embryo following fertilization.

A carrier of a balanced translocation has one normal copy of each of two chromosome pairs, and one copy with a translocation. Since one of each homologous pair is inherited, there are four possible outcomes for offspring. Left: a normal copy of each chromosome. Left-center: a balanced translocation. Right-center: one chromosome is normal, but the other has a translocation; this is an unbalanced translocation. Right: also an unbalanced translocation, but the other way round.

That completes our look at chromosomes. Let's now turn our attention back to the level of genes, and in particular to the different forms of these genes, the alleles. We've already considered dominant and recessive alleles and how they work, but now it's time to take a look at how these different alleles came to exist: the process of genetic mutation. Some of our genes will carry mutations that we inherited from one or both of our parents, which can lead to diseases that are passed down to later generations. In other cases, a new mutation occurs in a gamete, which leads to a disease in the offspring that was not present in either parent.

The A, C, G and T nucleotides that make up our DNA are functionally divided into groups of three consecutive nucleotides on the DNA strand. Since there are four different nucleotide letters, there are 4³ = 64 possible combinations of 3-letter 'words', called codons, and each codon makes one specific amino acid out of a total of 20 (this redundancy means that most of the amino acids can be coded for by several different codons, as shown in the table below).

A table showing which 3-letter codons (upper case) code for which amino acids (three-letter abbreviations). Note the presence of start and stop codons. Note also that the nucleotides are labelled either A, C, G and U (rather than A, C, G and T). Don't let this confuse you; although A, C and G don't change between DNA and RNA, a letter T in DNA is equivalent to a U in RNA, and the table shows codons that have already been transcribed from DNA to RNA form.

Remember that most nucleotides within a strand of DNA are functionless, so cells need a way to find the islands of genes in the much larger empty sea surrounding them. This is accomplished via 'start' and 'stop' codons: one particular codon is used to indicate the start of a gene and another (actually, one of three variations) is used to indicate the end.³

Because each specific three-letter codon codes for a specific amino acid, if there is a mutation (if one of the nucleotides is changed to a different nucleotide) this will often cause the codon to code for a different amino acid, and thus the resultant protein that is made will also often be different. The proteins that are built up from amino acids are very complex macromolecules that each form via an intricate and convoluted folding process; protein folding is highly sensitive to changes in amino acid sequence.

Protein folding. On the left is a sequence of connected amino acids, formed from the translation of RNA into protein. This sequence undergoes a complex biochemical folding process, based on the specific properties of each amino acid in the sequence. This results in a final folded protein of specific shape and function. Any changes to the amino acid sequence will usually affect the folding process and thus the final protein product.

Each of the 20 amino acids have differing chemical properties, so if one amino acid is replaced by another in a sequence, this can have a dramatic effect upon folding and thus on the final protein produced. This type of mutation is called a missense or point mutation, since the mutation occurs in a single nucleotide and leads to the production of one different amino acid in the protein. The effects of this can be large or small, depending on the specific amino acid change and its location in the protein.

Missense mutations. At the top we see how the nucleotide sequence of a gene forms 3-letter codons, each coding for a specific amino acid. The amino acids are collectively aligned into a protein of a given sequence. Below we see a point mutation in the fourth codon has changed the original letter A to a letter C, resulting in a change of codon from CAT to CCT, which in turn changes the amino acid from His to Pro. Missense mutations like this will often affect protein folding and function.

Another type of mutation whose effects are always large is called a nonsense or stop mutation. This occurs when the mutation causes the codon to change from coding for one of the twenty amino acids to coding for a 'stop' (one of the 3 out of 64 codons that indicates the end of a gene sequence). This type of mutation prematurely terminates the transcription of the RNA sequence. As this usually occurs in the interior of a sequence, it results in the loss of many amino acids; a nonsense mutation therefore tends to destroy the entire function of a protein.

Nonsense mutations. Again we see an original sequence of codons at the top, with the same sequence following a point mutation below. However, in this case the codon is changed from CAG to TAG. From the table above we can see that the DNA codon TAG (transcribed to UAG in RNA) is a stop codon. So therefore the codon sequence is terminated at this point, and the remaining codons are not transcribed into RNA or translated into amino acids. This drastically truncates the protein, nearly always destroying its function.

If we consider the genome as a whole, most mutations are actually irrelevant, as most of the genome is insensitive to sequence. In sections of the genome that are sequence-sensitive (such as those sections that are part of genes), most mutations are harmful, and most of the rest are neutral or nearly neutral. That most mutations are harmful should be intuitively obvious from the aphorism if something isn't broken, don't try to fix it; any changes made to a functional object are most likely to break the object. Most of the ways to change an object that don't break it still won't improve it (this corresponds to neutral genetic changes). The few remaining changes are beneficial.⁴ It is these neutral or beneficial mutations that occasionally become established in a population as a new allele for a specific gene.

Genetic mutations have several causes: errors during DNA replication leading to the wrong nucleotide being placed into a daugher sequence; environmental factors (e.g. many chemicals and ultraviolet light) that damage DNA which is subsequently affected by errors during repair; or from the insertion or deletion of DNA segments (called transposons) that can move around the genome.

Mutations are the raw material of evolution as well as a cause of many diseases, but as we aren't going to discuss evolution in this series, here I'm just going to focus on a certain type of harmful mutation: those that can cause cancer.

Cancer is a genetic disease. Many people carry known cancer risk mutations and can pass these on to their children. All of us acquire genetic damage during the course of our lives, some of which is carcinogenic. Virtually everyone who lives to later adulthood develops multiple cancerous or precancerous cell clusters (tumours) in various parts of their bodies.

Cancer is simply uncontrolled cell division. You will remember from earlier in this series that as old cells age, they wear out and must be replaced. Cells replace themselves by dividing, which they do once per cell cycle. This process of cell division is normally tightly controlled; a cell must only be allowed to divide when appropriate, so most of the time the cell cycle is paused at one of several cell-cycle checkpoints.

The cell cycle. After a cell has divided, the daughter cell is either in G1, where the non-genetic cell contents are duplicated ready for the next cell division, or it exits the cell cycle by going into a non-duplicative state called G0. In S-phase, the chromosomes are duplicated ready for the next cell division. The cell performs quality checks on the duplicated genetic material in G2, making any needed repairs. Mitosis and cytokinesis are the actual processes of cell division (discussed previously). Note that G1, G2 and Mitosis all have checkpoints associated with them, which halt the cell cycle until various necessary processes have been carried out.

This regulation of the cell cycle is controlled by genes called cell-cycle inhibitors, the most well-known of which is a tumour suppressor called p53. This protein works to repair damaged (mutated) DNA, and can also pause the cell cycle to allow damaged cells to be repaired or, failing this, initiate cell death in unrepairable cells.

Like most genes, the one that codes for p53 is part of a genetic pathway, in which it interacts with other genes via proteins and RNA molecules. If the gene for p53 gets damaged, the entire pathway of which it is a part is affected, and p53s vital functions as a tumour suppressor are eliminated, which makes the development of continuous, uncontrolled cell division much more likely. Most human tumours contain a mutation of the p53 gene.

Usually, multiple carcinogenic mutations are required before an individual will go on to develop cancer. Luck plays a big role in this, but of course we'd also like to minimize mutation rates as much as possible, and so reducing exposure to known carcinogens can go a long way in preventing cancer.

A common approach to cancer treatment is specifically to attack those cells undergoing uncontrolled cell division (i.e., tumour cells). The two usual methods of doing this are chemotherapy and radiotherapy. Tumour cells can be specifically targeted precisely because they are undergoing rapid cell division; recall that cell division requires the genome to be copied, and for this to occur the DNA must unwind to allow access by the replicative protein machinery. In this state the DNA is most exposed to damage, whether via chemicals or radiation.

But mutations aren't always bad; we can also make them work for us, even if we don't cause the mutations ourselves. One example of this is selective breeding, which we have been doing for millennia. Selective breeding is a way to get organisms to have particular traits that we find agreeable. All domesticated animals as well as all our food crops have undergone extensive selective breeding. For example, all modern dog breeds were developed by selecting for mating just those animals with the desired traits for the particular breed standard being aimed at. Every dog currently existing is descended over thousands of years from a species of wild wolf, and is a cousin of modern grey wolves.

All dog breeds descend from wild wolf ancestors. Traits such as docility, loyalty, and tameness were selectively bred into the lineages over many generations.

Plants have also been selectively bred for thousands of years. Originally, wild plants were cultivated and those specimens with the best traits (yield, and drought-, insect- and disease-resistance) were picked for reproduction whilst less desirable specimens were eliminated by preventing them from reproducing.

Broccoli, cabbage, turnip, kale and cauliflower were all bred from a particular species of wild mustard. The various traits of each plant type were selectively bred for, as shown.

Selective breeding can take dozens of generations to fully incorporate desired characteristics. It's also prone to various problems, such as inbreeding. Inbreeding results from the repeated breeding of closely-related individuals over several generations, and leads to increasing loss of genetic variability as each generation goes by. This makes organisms more susceptible to various diseases, due to harmful recessive mutations getting fixed in a population, and also leaves the species as a whole less able to adapt to new environmental challenges that arise.

In recent years, geneticists have developed the ability to directly modify specific genes, which greatly speeds up the process of improving species of interest, as well as potentially avoiding the inbreeding problems associated with traditional selective breeding. Genetically modifying organisms in this way remains controversial, not least because of the abruptness of the change, and also because genes are frequently transferred across species. The ethical implications of this need not concern us here.

Genetics is now a high-technology discipline, and research is intense in many fields. In a future article, I'll give you an overview of some of this current work, along with informed speculation on some of the directions in which the field may go over the coming years.

Let's now conclude this series on genetics with a case study, bringing together many of the concepts we've covered over the three articles.

1 out of every 2 million people are unlucky enough to be born with a genetic mutation that leads to a condition called fibrodysplasia ossificans progressiva (FOP). As with several rare bone-related disorders, this is a particularly nasty disease, characterized by the progressive replacement of connective tissue (muscles, tendons and ligaments) by bone tissue. This process starts with the neck and shoulders, working its way down the body. Subsequent injuries and infections accelerate disease progression, often leading to the complete seizing-up of joints and a gradual and irreversible loss of mobility.

FOP is an autosomal dominant disease caused by a sporadically recurring mutation in a gene called ACVR1, which is located on chromosome 2. This mutation usually occurs de novo (i.e., it originates in the foetus, rather than being inherited from a parent), and leads to a change in a single amino acid. FOP is usually⁵ caused by a mutation in codon 206 of ACVR1, where a change from G to A leads to the production of the amino acid histidine in place of the similar amino acid arginine. This simple event, occurring in one parent's gamete and thus being present in all the cells of the offspring's body, is all that is necessary to cause FOP and the many severe symptoms of the disease.

There is a specific class of protein that receives chemical signals originating outside the cell, called protein receptors, which respond to these signals by modifying some activity of the cell. ACVR1 is part of a family of protein receptors known as bone morphogeneic proteins, type 1 (BMP1). BMPs are cellular growth factors, and the BMP1 family are involved in bone and cartilage development. ACVR1 codes for a protein that is located in the body cells' plasma membrane and acts as a signal transducer, facilitating the transfer of information between the inside and outside of the body's cells.

Signal transducers, such as the protein coded for by the ACVR1 gene, are located in the plasma membrane of the cell. They can transfer signals from outside the cell to inside the cell via binding with messenger molecules (ligands). They then pass the signal into the cell using internal messengers, which trigger a chain reaction that eventually results in changes in gene expression.

The mutation in ACVR1 is of a type known to geneticists as gain-of-function: a class of mutations that result in the body producing a protein that has an abnormal function. In the case of FOP, the amino acid change results in the ACVR1 protein having a modified physical shape. Normally, protein receptors are inactive until they are activated by another specific protein binding to them (see above figure). However, the modified shape of the mutated ACVR1 protein leads to the receptor becoming active without requiring interaction with its normal binding partner. This increases its sensitivity, meaning that the gene is continuously activated at a low level, and is also much more prone to being fully expressed at inappropriate times.

Although the ACVR1 mutation must be present to cause FOP, it seems that this alone is not sufficient to cause the disease. Geneticists use the concept of expressivity to describe observed variability in the type, number and severity of the symptoms associated with a disease caused by a specific mutation. FOP exhibits variable expressivity: fairly large differences in the frequency and severity of flare-ups and disease progression between patients. Environmental circumstances (for example bodily injuries and infections) clearly play a role, but secondary mutations in the same molecular pathway could also have an effect, as could differences in immune system function and in the immediate environment of the soft tissue that is affected. These are questions that researchers are actively working on.

That concludes this three-part introduction to genetics. In future articles, I'll build on this background knowledge to discuss more specific aspects of genetics and how these relate to biology as a whole.

1. Chromosome 21 has the fewest genes of all the autosomes, making trisomy 21 the mildest of the trisomies. The next two chromosomes with the fewest genes are 18 and 13. Trisomy 13 (Patau syndrome) and trisomy 18 (Edwards syndrome) foetuses can survive to birth, but do not normally live more than a few months at most. Trisomies of other autosomes are lethal in utero.

2. Although we learned in part two that most genes are haplosufficient (a single functional copy is enough to produce the required amount of protein), a significant fraction of genes are haplo-insufficient (and so require both copies to be present and functional). As even the smallest chromosomes have hundreds of genes, this means that a monosomy will affect scores of genes that are haploinsufficient.

3. Actually, the genes themselves are divided into alternating sections called introns and exons. The exons are collectively spliced together when building an RNA strand representing an entire gene, while the introns are discarded. Exons can be spliced together in different combinations, so that a single gene can actually code for several different proteins.

4. At some future date I intend to write several articles on evolution, the unifying principle of all of biology. It's enough to say here that those few mutations that turn out to be beneficial have done a lot of the heavy lifting in building organisms over huge time spans.

5. There are a few reported cases where a different mutation in the same gene has occurred.

#genetics #series

This is the second article in a three-part series on genetics. In part one, we began with some basic concepts: looking at cells, the structure and replication of DNA, genes and what they do, and how cell division and genetic inheritance work. This article will delve a little deeper into some slightly more challenging concepts.

Let's go back to the genes. We have two copies of each gene, one from each parent. But what if these copies are different from one another? How does this change things? In fact, this is often the case. Most genes come in several different forms, called alleles. If genes were cars, then alleles could be Honda, Tesla or Lamborghini. Although affecting the same trait or traits, different alleles have differing DNA sequences, resulting in variations in the proteins they produce. Each of our 20,000 genes is located in a specific position (called a locus) on one of our chromosomes. All the alleles for a given gene can thus be found in the same position on the same chromosome as one another, but there can only be one of the possible alleles on any particular chromosome.

Now that we know what alleles are, let's look at the role they play in some common conditions.

Perhaps the simplest example in humans is albinism – a disorder resulting in a lack of melanin pigment in the skin, eyes and hair. One of the alleles for a gene that makes the protein required for melanin production codes for a defective version of the protein. Individuals who inherit this allele from both parents thus have no means of producing melanin and are albino. However, an individual with one defective allele and one good allele is still able to produce enough melanin to appear normal. So in this case, the good allele masks the defective allele, and geneticists say that the good allele is dominant, while the defective allele is recessive.¹ Albinism is thus a recessive condition: it only shows up in people in whom both alleles are defective.

Remember that each parent gives one of their two alleles to each of their children, and that since this is determined at random, there's a ½ (50%) chance for a child to receive either allele (see figure, below). So, if two albinos have children, each child will also be albino (as the child will inevitably receive a defective allele from each parent). If an albino has a child with a non-albino, the albino parent will pass on one of their defective alleles to their child, but the non-albino will usually have two good alleles, and so the child will have one good allele and thus will be non-albino. However, note that in this case the child will carry one good allele and one defective allele, and so if the child subsequently had their own children and the other parent was albino, then each of these children would have a ½ chance of also being albino. And if two non-albino carriers (each having one good and one albino defective allele) had children, each child would have a ¼ chance of being albino (as shown in the figure below).

Autosomal recessive inheritance, for a trait such as albinism. In this case, both parents are carriers of the disease, so although they are themselves unaffected they can pass the trait on to their offspring. Note that if one parent was affected and one was a carrier, the affected parent would always pass on the disease allele, and half their children would be expected to be affected, based on which allele the carrier parent gave.

As well as recessive diseases (or non-disease traits), there are also diseases that are dominant. In these cases, the disease allele is dominant over the normal allele. This means that an individual will be affected by a disease even if they carry only one bad allele.

A generic example of autosomal dominance. Here, the father has a dominant disease allele and is affected. Each of his children has a 50% chance of receiving the disease allele and also being affected. Note that in this case the mother is free of disease alleles and so does not affect any of her children in relation to this particular disease (since she'll always pass on a healthy allele).

Dominant diseases tend to show up more frequently in a population than recessive diseases (because you only need one bad copy of the allele to be affected, rather than two, and so there are no unaffected carriers like there are with recessive diseases). This means that dominant diseases are generally either less severe than recessive diseases, or they show up later in life (otherwise these alleles would tend to eliminate themselves from a population). This is the case with Huntington disease, a lethal neurodegenerative disorder that is usually only clinically apparent at around age 40. By this age, many people will already have had children and may have passed on the Huntington allele to some or all of them. In this way, the Huntington allele does not eliminate itself from the population. Fortunately, as we'll consider in the final part of this series, there are now hundreds of diseases for which genetic screening is available, and individuals with a family history of such a disease can now be tested prior to having children.

What's actually going on at the molecular level that makes some traits dominant and some recessive? As we've learned, our genes produce proteins, and it turns out that a dominant allele is dominant because only one of the two alleles is required to produce sufficient protein. In the case of albinism, heterozygotes have a normal phenotype: although they have only a single functional allele, this is sufficient to produce enough protein to avoid albinism. We say that the normal allele is dominant over the albinism allele, because one normal allele produces enough protein to mask the effects of the albinism allele. On the other hand, in the case of Huntington's disease, it is the disease allele that is dominant over the normal allele. This is because two normal alleles are required to provide enough protein for normal function.

So far, we've considered only the effects of autosomal alleles. Things get more complicated when the gene is on the sex chromosomes. As we already learned, there are very few genes on the Y-chromosome, and those that are present are mostly involved in determining maleness. However, the X-chromosome does carry genes that code for many different traits. As with most genes, having a single correctly functioning gene on either of the two chromosomes is sufficient for normal function, so females carrying a disease allele will normally be unaffected. But remember, although females have two X-chromosomes, males have only one: if there is a mutation in an X-chromosome gene in a male, there is no 'backup' copy of the gene and so there will often be incorrect or non-existent functionality for the trait affected by that gene.

One familiar example of this is red/green colourblindness. The structures that allow us to perceive colours are the eye's retinal cones. Humans have three different types of cones, and the genes that control their development are on the X-chromosome. The different cone cells each contain a pigment sensitive to specific wavelengths of light, peaking in sensitivity for either red, green or blue light. Between them, the three types of cone cells cover the entire visible light spectrum, and can be used in combination, allowing us to distinguish millions of different colours. A mutation in one of the genes degrades or eliminates the ability of the person carrying the mutation to see the associated colour. In females, the effect of the mutation is negligible, since the second X-chromosome adds redundancy. However, in males the mutated gene is the only one of its kind, leading to the inability to see the colour in question. This is why colourblindness is much more common in males than in females.

X-linked inheritance, for a trait such as red/green colourblindness. The gene in question is located on the X-chromosome. Here, the father has one normal gene on his X (so is not colourblind), while the mother has one normal copy and one mutated copy on her two X-chromosomes. The mother herself is unaffected (a carrier) because the disorder is recessive, but her children each have a 50% chance of inheriting the mutation. If this happens to any of her daughters, they will also become carriers, while any sons who inherit the chromosome with the mutation will be colourblind.

X-linked traits like colourblindness tend to skip generations, and with a little thought we can understand why. Suppose we have a normal-sighted female and a colourblind male, and this couple have children (see figure, below). The female parent has X-chromosomes with fully functional colour genes, so she won't contribute to colourblindness in any of her children. The male parent, however, being male has only one X-chromosome, and this chromosome has a non-functional colour-vision gene (which is why the male parent is colour-blind). The male children of this couple will receive their X-chromosome from their mother, and a Y-chromosome from their father, so their colour-vision will be normal. The female children will receive an X from both parents: a good X from their mother and a bad X from their father. Since one out of two good X-chromosomes is all that is required for normal colour vision, the female children (just like the male children) will all have normal colour vision. However, these female children will be carriers of colourblindness. Any children they subsequently have will have a 50% chance of receiving the X-chromosome carrying the colourblindness mutation, meaning that female children will also be carriers and male children will be colourblind. Hence, the colourblindness skipped from a male in the first generation to males in the third generation via one or more unaffected female carriers.

Read the previous paragraph through again if it wasn't quite clear.

X-linked recessive disorders are passed down generations via female intermediaries. Here, we have an affected father who gives his X-chromosome containing the defective gene to all his daughters and none of his sons. His daughters, although themselves unaffected, can subsequently pass on the defective gene to their own children, and their male children who inherit the defective gene will be affected.

There are also X-linked dominant disorders. These are all rare and beyond the scope of this article.

As we know, genes operate by producing proteins, and the amount of protein produced depends on not only the specific alleles present, but also on the specific type of cell and which part of the body the cell is in. However, gene expression also varies over time: a developing human will have greatly differing gene expression as a foetus than it will as an infant, child or adult. Human development and gene expression are extremely large and complicated subjects, and we won't go into them in much detail here; let's just settle for a brief overview.

An example is illustrated by the enzyme lactase. Enzymes are proteins that catalyze (accelerate) reactions that occur in the body. Lactase is the enzyme that allows milk sugar (lactose) to be properly digested; without it, we develop the symptoms of lactose intolerance.

In nature, milk is only produced by nursing mothers (and milk production is one of the defining characteristics of mammals). Once an offspring is weaned, its production of lactase is greatly reduced, since there is no longer a need for it. However, with the advent of cattle domestication, some humans were able to produce non-human milk products as food on a regular basis and consume these throughout their lives. It turns out that individual humans have differing levels of lactase activity, and much of this difference can be explained by genetic variability: selective pressures in post-agricultural human populations led to the spread of a particular dominant mutation through those populations utilizing dairy production.² This mutation causes lactase persistence – the ability to digest lactose beyond weaning due to high levels of lactase in the small intestine. Other populations, who didn't domesticate cattle, continue to down-regulate lactase after weaning, resulting in lactose intolerance. Prior to weaning, there is little difference in lactase expression between individuals, whether they have the mutation or not. However, in those not carrying a lactase-persistence mutation, regulatory proteins are produced following weaning that can bind to and interfere with the lactase gene, greatly reducing the gene's expression and thus leading to lactose intolerance.

A map showing 'old world' lactase persistence frequencies. Lactase persistence mutations arose independently in Europe, sub-Saharan Africa and the Middle East, all of which regions developed dairy farming within the last few thousand years. As populations became more mobile and spread to various other parts of the world, they took the alleles with them.

At the molecular level, genes are transcribed into mRNA via proteins that bind to a section of the DNA called the promoter. Other proteins (called polymerases) open up the DNA double helix and move along it, transcribing each letter of DNA into the corresponding letter of RNA. If a competing protein binds to the promoter before transcription can begin, this shuts down the expression of the gene. Thus, a cell can begin transcription by allowing an enhancer to bind to the DNA, or prevent transcription by allowing a repressor to bind. In those individuals exhibiting lactase persistence, the repressor protein for the lactase gene is not produced, which allows the continued production of lactase.

Other genes are regulated according to environmental conditions. For instance, if you ingest a poison, such as alcohol, your body responds by upregulating certain genes to metabolize the substance (in the case of alcohol, these include alcohol dehydrogenase, aldehyde dehydrogenase and fatty acid ethyl ester synthase, mainly in the liver and pancreas).

As well as affecting the amount of mRNA produced, there are many other forms of regulation: some involve modifying the way the mRNA is subsequently processed; some modify how much mRNA is allowed to exit the nucleus prior to translation into protein; others control the rate of translation into protein or the rate at which protein is degraded (see figure). The important point to bear in mind is that all these regulatory mechanisms serve to fine-tune the functioning of an organism based on a specific set of conditions.

There are many stages at which protein production can be regulated. Starting inside the nucleus, cells can regulate whether, and how much, DNA is transcribed into mRNA; how mRNA is modified before leaving the nucleus; how much mRNA is allowed to leave the nucleus; whether and how many times a strand of mRNA is translated into protein; and how long a particular protein lasts until it is degraded.

So far, we've thought of genes as acting independently of each other, but in fact this is rarely, if ever, the case.

We've just seen that the lactase gene is regulated by proteins that are produced by other genes, and which serve either to increase or decrease the gene's activity. But even this has greatly oversimplified what happens in the majority of cases. Take eye colour, which isn't simply determined by a single allele for each possible colour. Remember from an earlier example that pigmentation in eyes, hair and skin depends on a protein called melanin, which is produced by specialized body cells called melanocytes. Melanocytes are highly variable in the amount of melanin they can store, and this in turn leads to variability in the colour of eyes, hair and skin in individual humans. Eye colour is determined by the amount of melanin in the iris. The greater the amount of melanin, the more light is absorbed by the iris, and the darker the iris appears; blue eyes have low levels of pigment, whilst brown eyes have high levels.

Not only are there genes that affect the production of melanin, but also genes that affect its processing, storage and transport. In all, there are over one hundred genes that influence human pigmentation, at least eight of which are known to affect eye colour. Interactions between these genes all influence the amount of melanin that is taken up in the iris, and thus combine to determine eye colour. Furthermore, most of these genes will affect multiple other traits as well. This is generally the case; there are few genes that only affect a single trait. Rather, genes tend to work in overlapping networks, and their effects cascade down through genetic pathways and are in turn affected by other genes and feedback mechanisms.

The story gets yet more complicated when we consider the effects of the environment. The old argument of nature versus nurture to a large extent involves both sides arguing past one another. As we can probably appreciate, the genes that direct the building and functioning of our bodies are themselves affected by the outside world. It should be intuitive, for example, that the quality of nutrition that a child receives will affect the child's growth and development, both physical and intellectual. Diseases and infections also affect growth.³ There are many hundreds of genes that affect human growth, which not only interact with each other in large genetic networks, but are also affected by outside forces. Likewise, as we are all familiar with, melanin production is affected by light intensity. If you move to a high altitude, your body gradually increases the number of oxygen-carrying red blood cells, to counteract the lower amount of oxygen that is in the atmosphere. There are countless other examples we could enumerate.

So we have now moved far beyond our initial, overly-simplistic, view of genetics as one gene —> one trait. We know that most genes have many different versions, called alleles, and that some alleles are dominant over others. We also know that things work slightly differently when considering the sex chromosomes rather than autosomes, that gene regulation can give endless nuance with regard to the expression of genes and the proteins that are produced, that genes tend to have multiple effects, and that they interact with other genes and the environment.

In the final part of the series we're going to look at some topics in applied genetics, including chromosome variations, genetic mutation, cancer, selective breeding and the genetics of diseases.

1. Note that although we're looking at diseases here, the concepts of dominance and recessivity also apply to traits generally, some of which are benign variation rather than diseases.

2. Interestingly, the specific mutation varies by population. European, East African and North African populations each have a different lactase persistence allele.

3. Improvements to general levels of nutrition and health care over the last century or so in developed nations have led to dramatic increases in human growth, noticeable even from one generation to the next.

#genetics #series

Why do some traits appear in grandparents and grandchildren, skipping the intervening generation? Are there really genes for intelligence, running, mental arithmetic? More broadly, are we in any sense controlled by our genes? How far has our understanding of genetics advanced in recent years, and where is the field heading? What about ethical issues raised by new genetic technologies?

A conceptual understanding of genetics is mandatory for anyone wishing to truly understand questions like these. And, if we don't want to be misled by grandiose or unfounded claims that are regularly made about genetics, we need a solid grasp of how genetics works. This is today more vital than ever, given the ever-accelerating pace of research.

This is the first of a three-part series that collectively aims to cover all the major concepts required for a solid understanding of modern genetics. This article will give you a tour of the basics, with no prior knowledge required. In part two I'll expand on this foundation by covering more complex concepts, and part three will conclude the series by discussing some human-specific areas of genetics in greater detail.

After reading this series, you should have a much more sophisticated understanding of genetics, and you'll be able to utilize this as I plan to write more specialized articles on diseases and other genetics-related stories in the news, several of which will be critiques of a few of the many dubious claims propagated by the media.

Let's start the discussion with three somewhat familiar concepts: cells, genes and DNA.

Living things are all made up of cells, self-contained biological units consisting of various components that are themselves made up of nucleic acids, proteins, carbohydrates and fats, all enclosed in a porous membrane. Whether part of an organism consisting of trillions of cells, like a human, or the entire organism as with bacteria and yeast, the cell is the fundamental building block of life.

An animal consists of many different cell types. Cells of a specific type often cluster together to form tissues. An organ (e.g., the skin or heart) consists of several tissue types. Finally, organ systems are composed of several organs working together (e.g., the nervous system consists of brain, spinal cord, the sense organs and nerves).

Although each cell type is optimized for its own specialized purposes, all animal cells consist of a selection of the same subunits, called organelles, each of which has a specific function (energy production, waste disposal, repair and growth, etc). The cell membrane allows adhesion to neighbouring cells and also communication with the outside world. Cells can receive instructions from outside via specific signalling molecules, electrical impulses, or changes in the conditions within the cell (for example acidity or the concentration of various molecules).

An animal cell consists of an outer membrane and many organelles, suspended in a liquid called cytoplasm. The DNA is stored in the nucleus (the large organelle close to the left side of this cell).

In order to correctly carry out most of the things that it needs to do, a cell requires accurate instructions. These instructions are contained within the genes. Humans each carry around 20,000 genes in almost all of their 35 trillion body cells. The genes collectively contain the information needed to build our bodies and keep them alive.

DNA is the molecule that contains these genes, and it is combined with proteins that help to store and protect it. DNA is organized in such a way that it acts as a code for the production of proteins. It's a very elegant molecule; the code is made entirely from just four different small subunits called nucleotides, connected into a very long, very thin strand of DNA. The nucleotides from one strand pair with those from another strand, to form the familiar double helix shape (see figure). Most of these nucleotide pairs don't do much of anything, but strewn amongst the dead viruses and repetitive sections of junk are the genes, and the switches that control them. A typical gene is a few thousand nucleotide pairs in length, consisting of a specific sequence of the four different nucleotide subunits. It is this sequence that determines which protein(s) the gene can make.

Left: DNA consists of four nucleotides, designated A, C, G and T. Notice that A always bonds with T, and C always bonds with G. Right: the double helix structure of DNA.

Because our bodies consist of many different types of tissues and organs (each with different functions), each body cell that makes up these tissues and organs needs different proteins at different times. So we can't just have all our genes busily making proteins around the clock; our genes need to be regulated – they need a way to produce certain proteins in certain amounts and at certain times. And there's a lot to keep track of: our genome – the entirety of our DNA – has over 3 billion nucleotide pairs. You could fit about 30 million pairs into 1 centimetre (75 million pairs per inch). If you stretched out the DNA in one of your cells, it would be around 2 metres (6 feet) long. Multiply this by the number of cells in your whole body, and you'd have enough DNA to stretch to the Sun and back over 50 times! And if we multiply that by the 7 billion people on the planet, then we'd get enough human DNA to stretch out of our solar system and past all the stars we can see in the night sky. In fact, all this DNA would be able to leave our own galaxy and reach into a neighbouring one.¹

How do our bodies take care of and organize all this DNA? Within each body cell a genome is packed into a central structure called the nucleus (see figure). The sheer volume of genetic material, combined with the necessity that most genes be switched off most of the time, means that the usual state of our DNA is to be kept tightly packed, folded, rolled and scrunched up. This is accomplished using proteins, in a DNA-protein complex called chromatin. The chromatin is kept compact and stops the genes from making new proteins when they shouldn't. On those occasions when a gene is called upon to go to work, the section of chromatin containing the gene in question is unfurled, unpacked, unwound and unscrunched long enough for it to make the requisite amount of protein, after which it's bundled back up and packed away like an insane granny in the attic.

The nucleus is the largest organelle in the cell. The nuclear envelope has pores to allow the transfer of particles into and out of the nucleus. The chromatin consists of DNA wrapped in various proteins. The nucleolus is a genomic region involved in the production of ribosomes, small cellular machines that reside outside the nucleus and convert genetic information into proteins.

As well as being tightly packed, our DNA (in the form of chromatin) is also divided into 23 pairs of chromosomes. We inherit one of each pair from each of our parents, giving us two of each chromosome. For example, a child's mother has two copies of chromosome 1, as does their father. This child got one of their mother's chromosome 1, and one of their father's. The first 22 pairs of chromosomes are known as autosomes, and thus we have two copies of genes on any of our autosomes – one from each of our parents.

Inactive chromatin is highly condensed and divided into 23 pairs of chromosomes.

The 23^rd pair of chromosomes is different; these are the sex chromosomes. The most important function of the sex chromosomes is – you guessed it – determining a person's sex. Females have two X chromosomes, while males have an X and a Y. Everyone inherits one of their mother's two X chromosomes, but girls inherit a second X, from their father, while boys inherit their father's Y chromosome. It is thus the father who determines the sex of his children. A runt among chromosomes (see figure), the Y has very few genes, and those that it does have are almost exclusively devoted to producing maleness in their owners.

The 23 pairs of human chromosomes, imaged in a karyogram. These are from a male (you can see that the 23^rd pair of chromosomes (which are the sex chromosomes) consists of an X and a Y).

So now we know how our genetic material is organized and that the genes make proteins. We also know that a gene consists of a specific sequence of nucleotides that determines the protein it will produce, but how does this work? As we learned already, the chromosomes are safely stored in a central, cordoned-off area of a cell called the nucleus. When a cell needs to make a protein, it duplicates the relevant section of DNA by copying the sequence into a molecule called messenger RNA (mRNA). mRNA is single-stranded and disposable, and it moves out of the nucleus through pores in the nuclear envelope, where its code is read and interpreted by ribosomes that produce the required protein, after which the mRNA is recycled. The DNA is like the original, master copy of a vital document, and so is kept safe inside the nucleus. mRNA is a transcribed copy of a specific section of the master document, and so can be removed, read and thrown away as required.

But how does the cell control which genes are operative at a given time? In addition to the genes, and usually near to them, DNA also contains sequences of nucleotides that act as switches, which can be turned on or kept off by specific regulatory molecules (proteins or regulatory RNA strands). This regulation can occur in one or more of several different ways: a cell can keep the DNA tightly packed and inactive, or open it up for transcription using structural or chemical signals; it can control the timing and duration of mRNA production from opened DNA using regulatory switches and proteins (or even other types of RNA); it can modify transcribed mRNA, perhaps rearranging it to form different sequences and so make different proteins; and it can send signals to increase or decrease the number of copies of a specific mRNA molecule, to determine how much protein an mRNA produces.

All these mechanisms (and more which we won't consider here) work in concert to regulate protein production with great precision. In this way, each cell in our body can produce different proteins at different times, as its circumstances dictate.

There's endless wear and tear involved in being alive, and up until adulthood our bodies are constantly growing. Even as adults, throughout our bodies cells are continuously dying or being killed, and must be replaced. Around 2 million red blood cells die per second; if you've ever donated a pint of blood, you lost around 2.5 billion red blood cells (plus many other types of cells). All our other body cells also wear out and die, and can also succumb to injury. A child starts out as a single fertilized egg cell, and within a couple of decades has become an adult of 35 trillion cells. All this means that large multicellular creatures like us must produce a huge number of new cells each day.

A new cell comes from an existing cell that divides into two. We learned earlier that our cells each carry a copy of our entire genome. So before a cell can divide into two new cells (and in addition to replicating all its other components), it must make an extra copy of the genome so that both daughter cells can each have one. This requires the unwinding and copying of every one of the 3 billion nucleotide pairs. Because DNA is double-stranded, it can be separated into its two constituent strands. Each single strand can then be used as a template to make the complementary strand, producing two identical double-stranded daughter molecules of DNA.

How does this work? Recall from an earlier figure that there are four nucleotides (A, C, G and T), and that an A from one strand always bonds to a T from the complementary strand, and likewise C always bonds with G. This means that when the cell unwinds the DNA double helix, it can look at each single strand of DNA and will always know how to build the other, complementary, strand. For example, if the next nucleotide on a template strand is an A, the cell will add a T to the complementary strand. This elegant process of DNA replication, though it has amazing fidelity, is still prone to the occasional mistake; these mistakes are called mutations, and we'll talk about these and their implications another time.

Cell division (simplified only to include one pair of chromosomes). Before cell division can occur, each chromosome must duplicate itself into two sisters, connected to each other near the centre, as has already occurred here (Prophase). The chromosomes are then aligned along the central plane of the cell (Metaphase), using tiny tubes. The sisters then move down the tubes to opposite poles of the cell (Anaphase), after which two new nuclei form around the separated chromosomes (Telophase) and the cell physically divides in two new daughter cells, with one sister from each pair in each (Cytokinesis).

We know that a human starts out as a single cell that got one copy of each chromosome from each parent. If you think about this for a second, you'll realize that what we've discussed so far concerning cell division won't work for sexual reproduction: if sex cells had the same amount of genetic information as other body cells, an offspring would end up with twice the amount of DNA as its parents, and this doubling of DNA would compound in each generation.

So an egg cell and a sperm cell must have half the number of chromosomes of our other body cells (i.e., a single copy of each chromosome, rather than a pair of each). How does this halving work? Normally, each chromosome is copied before cell division occurs, so the chromosome number doubles prior to cell division and is halved during cell division, leaving the original number of chromosomes in the daughter cells.

But during cell division for the production of eggs and sperm (a process known as meiosis), this duplication is followed by two rounds of cell division, leading to an overall halving of the number of chromosomes that end up in the egg or sperm cell.² Each chromosome is present as a singleton, rather than being one of a pair. So when a sperm cell fuses with an egg, the single chromosomes contributed by each cell combine into pairs, ensuring the embryo has the correct amount of DNA.

Egg/sperm production. Top: one of the 22 pairs of homologous autosomes in a cell, the red member of the pair inherited from one parent, the blue one from the other. Next: each chromosome of the pair is duplicated. Bottom half: two rounds of cell division first take the chromosome number back to normal, then finally produce four (egg or sperm) cells, each with half the number of chromosomes. When a sperm cell fuses with an egg, the genetic material combines to again produce a cell with a complete set of chromosomes.

Meiosis includes a couple of processes that serve to greatly increase genetic variation in offspring. These both occur during the first of the two rounds of cell division, after the chromosomes have duplicated. The first occurs due to the random inheritance of one parental chromosome from each of the 23 pairs. When the chromosomes line up prior to the cell dividing, the cell does not check which parental chromosome is on which side of the dividing line, leading to a randomization of chromosomes in the two daughter cells. This means that, on average, siblings will share (a random) half of their genetic material with each of their brothers or sisters.

If we had only 3 pairs of chromosomes, there would be 2³ = 2 x 2 x 2 = 8 ways to divide them between daughter cells. Here we see four different possible initial alignments along the central vertical axis, as used for cell division; you can see the eight different possible combinations that could occur (after the left-hand chromosomes have separated from the right-hand ones). Since we actually have 23 pairs of chromosomes, there are in fact 2²³ = 8.4 million possible arrangements.

The second process that increases genetic variation is a mixing-up of the alleles by swapping sections of chromosomes, known as crossing over (see figure). Because the two chromosomes are analogous to one another along their entire length (they contain different versions of the same genes at any given position), variability can be increased by breaking the chromosomes and exchanging one or more sections between them. The resulting unique chromosomes later go their separate ways, each ending up in a different gamete (egg or sperm cell).

Left: Each of the chromosomes inherited from a parent have been duplicated into a homologous pair (the second step in the above figure on egg/sperm production). Centre: two of the homologues exchange sections. Right: the resultant four chromosomes are divided between four gamete cells, each with differing genetic material.

Together, these two mechanisms can produce almost unlimited shuffling of genes, which is why siblings usually look similar, but can be strikingly different in certain characteristics.

In part two of the series, we'll consider the effect that variations in genes have on individuals, look at how some well-known traits work, and examine some more complex situations such as how the outside world interacts with our genes and the effect this has upon us.

1. We won't go into the calculations that include all living things on Earth; it's sufficient to state that all that DNA would comfortably stretch around the entire known universe dozens of times.

2. There is an initial doubling of the genome, followed by two halvings, thus leaving half the original, undoubled, amount of DNA (1 —> 2 —> 1 —> 0.5).

#genetics #series

The form of poker I'm going to focus on is No Limit Texas Hold'Em (NLHE), which has long since eclipsed the other forms of the game in popularity. Note: I'm not going to go over the rules of NLHE here; if you need to, you can read the Wikipedia article and then come back.

NLHE's dominance is justified, since it demands the greatest amount of decision-making, and thus offers skilled players the most opportunities to win. Let's think about this for a minute. If you've ever watched an old Western movie, you'll doubtless have seen a bunch of n'er-do-wells sitting around a table in the saloon, playing poker. What kind of poker? Invariably, it's Five Card Draw. Everyone gets five cards, face down, there's a round of betting, everyone gets to make a single exchange of cards, there's a second round of betting, followed by the showdown (the kind with cards, and possibly also the kind with six-shooters). But this isn't very interesting; the only information available to a player concerning what their opponents may be holding is based on the opponents' betting and how many cards they exchange. That's better than nothing, and certainly a skilled draw player can make some fairly accurate inferences (especially if their opponents are well known to them), but this lack of information places a definite limit on the skill involved. Furthermore, with only two betting rounds, there's not even much scope for bluffing.

Now consider a game of no limit hold'em. Each player gets two hole cards, face down. There's a betting round. Now three community cards are dealt face up. Another betting round. A fourth community card, and another betting round. Then a final community card and a final betting round. In this situation, there's a huge amount of information available to an observant and skilled player. Each player gets four rounds to fold, call, bet, raise or re-raise. At least three of their cards are known to everyone else. Did they start betting after a third heart hit? Or when an ace came? Did they raise early but then back off when the ace hit? Have they been raising aggressively on almost every hand for the last couple of rounds? Did the most timid player at the table suddenly stifle a grin and reach for his chips?

As you can already see, there's a lot of information to take in. But so far we've only scratched the surface. So, to help make things clearer, let's go through a hand of NLHE. We'll run it twice, firstly as a novice might play it, then as a good player might. There will be plenty of digressions along the way to explain the many concepts involved.

Imagine we're playing a home game of NLHE. We've been playing for half an hour, and there are five other players. The game is 5¢/10¢ blinds, and everyone has $10-$15 on the table. We are in the dealer position (button), and after we deal the hole cards, we look down at a pair of black aces (AcAs). Perfect! We have the best possible starting hand. An annoying player who seems to win every week is first to speak (we'll call this early position, or EP for short), and he opens for 40¢. We try to keep calm: we've got him and now we can take some money off him, and maybe wipe the smug expression off his face. Even better, after the second player (mid position, MP) folds, the third player (cutoff) calls. More money coming our way! Now, we don't want to scare anyone off, so we don't raise yet, but instead cooly call. The small blind (SB) folds, and the big blind (BB) calls.

The positions in our six-handed poker game, with Player 6 dealing. The names for the positions are: 1 = small blind (SB), 2 = big blind (BB), 3 = early position (EP), 4 = mid position (MP), 5 = cutoff, 6 = button.

So there are four players left in the hand. We've had junk cards so far this evening, which have cost us money, so here we want to make as much back as possible. We deal the flop:

Kc 10h 4s

This seems innocuous enough, right? In fact, anyone who has a king has now made a worse pair than us, so they might stick around long enough for us to take some more money off them. The big blind checks, then the annoying guy (EP) bets $1.20. The cutoff folds. We want to string EP along, so we just call again. He's got no idea what's about to happen to him! The big blind folds. Two players left. We deal the turn card:

2c

A nothing card. It'd be nice to get another ace, but no bother. EP checks. He must've been bluffing, but we're not letting him off that easy! We bet $1.00 more, just milking him. He calls. Excellent; we might be able to get even more off him after the river, which is:

9d

EP looks at us, thinks for a few seconds, then moves $3.50 toward the middle! We almost beat him to the pot with our call. He flips over QhJh. Unbelievable, a straight on the river! No wonder this guy always wins; he's so lucky. And we didn't do anything wrong, did we? It's just more bad luck.

Actually, we made several mistakes. Let's discuss them to get an idea of how we could have played the hand better.

The same setup as before. Six players, 10¢ big blind, everyone has 100-150 big blinds. We deal ourselves a pair of black aces. EP open raises for 40¢, there's a fold from MP then a call from the cutoff and it's our turn to speak.

Firstly, let's be clear that we love our hand and we love that the pot was open-raised (even though the raiser is the best player at the table). AA is the best starting hand in hold'em and we can make some money here, but it's also only a starting hand and it is definitely vulnerable – you aren't entitled to win the pot just because you hold aces. The key point here is that AA play better against fewer opponents. Against a single opponent, you're around a 4-1 favourite to win the hand if it goes to showdown, no matter what hole cards your opponent has been dealt. Against three opponents with random hole cards, you're less than a 2-1 favourite, and against five opponents it's a coin flip. The more opponents we have, the greater the chance that someone will get lucky and hit a big hand that beats us. So we'd like to play our premium starting hands, amongst which aces are the best, against fewer opponents.

This is a good time to talk about starting hand strength. What hole cards should we be looking to play? The principle considerations here are the number of players and their playing styles, stack size (i.e., how much money each player has on the table, available to bet) in relation to the blinds, and our position at the table in relation to the button (dealer). Few casual players ever even consider the effect that position plays, but in fact position is one of the most crucial aspects of poker. Quite simply, the later you play in a hand, the more of an advantage you possess. Pre-flop, the big blind is last to speak and so has an advantage. However, in every betting round after the flop the dealer is last to speak and the blinds are in early position; they're at a distinct disadvantage. It's better to play last because you've got the most information – you've just seen what the other players have done and thus are better equipped to decide what to do yourself. So clearly, given the advantage that late position confers, our range of playable hands will be largest when we're the dealer, and smallest when we're first to speak. A small range consists of high-quality hole cards (big pairs, high cards like ace-king), while a large range also contains weaker hole cards (small pairs, weak aces like ace-nine, weak suited cards like king-ten of hearts).

Let's look at a specific example to see the importance of position. We might open (raise the big blind) from EP with a pair of eights, but then we'd have to wait to see what the other players do. If we get re-raised by the button, we either have to fold and lose the money we just bet, or continue with a marginal holding against a player who has position on us (acts later than us). Contrast this with the situation when we're in the big blind with 88, EP opens and the button re-raises. In this case, we know that our 88 are very likely no good, since two other players who both have position on us after the flop have shown strength and so can probably beat us. We can utilize this extra information, available because we're last to speak, to fold and thus not lose any money on the hand (other than the blind which no longer belongs to us anyway). This advantage of having good position is why everyone gets a turn at being the dealer, so that in each round everyone plays from each position at the table.

The next consideration when determining our playable starting hand range is the number of players and their playing styles. The first of these is simple; the fewer opponents we have, the larger our range of starting hands should be (because the average hand strength is relatively low). Playing style is a huge topic, which we'll consider in greater detail later on, but for now just consider its two main axes: loose—tight and passive—aggressive. The first concerns the proportion of hands a player will play. If you play lots of hands, you're loose, whereas if you play few hands you're tight. The second continuum refers to how you play those hands: if you tend to check and call a lot, you're passive; if you tend to bet and raise a lot you're aggressive. A reliable rule of thumb is to play the style opposite to your opponents' style; if most hands have many players calling many hands, the table is loose-passive and you want to play tight-aggressive (play a narrow range of hole cards, wait until the flop gives you a big hand, then keep betting and raising to get paid off by the passive players, who tend to call a lot with marginal hands). Most casual players are loose-passive, and are known as fish (presumably because they like to fish for cards, or maybe because they're easy to catch). Fish are also known as ATMs, since they'll give you money. The golden rule for playing against fish is you do not bluff; wait for a good hand and then bet it strongly (as they'll often call you with a worse hand). Another, less common, home-game type is the maniac. These guys like mindlessly to bet, raise and re-raise on almost every hand. Although it requires more nerve to play them, they're actually even easier to beat than the fish: since they're excessively loose-aggressive you play tight-passive against them. Wait to hit a big hand then let them bluff off all their chips to you (and don't forget to say thanks afterwards). There are many other playing styles, but minor variations on these two fundamental types account for the majority of players you'll find in a home game.

The other consideration for determining a hand range is relative stack sizes (a player's stack is all the chips they have in front of them). If the players are deep stacked (they have more than 100 big blinds), the value of speculative hole cards (e.g., suited connectors like 9d8d, suited aces, and small pairs) increases. This is because speculative hole cards, as their name implies, don't make a big hand very often, and most of the time you'll end up having to fold. This leads to losing many small pots. So on those few occasions when you do get a big hand with your speculative hole cards (a small pair becoming three-of-a-kind (a set), or connectors making a straight or flush), you'll need to win a big pot to be profitable overall. For this to be able to happen, you and your opponents must each have a deep stack of chips. Conversely, if stack sizes are small (e.g., around 50 BB), speculative hands have little value and won't pay off over time, as you won't be able to win enough on the hands you hit to cover your losses from the majority of hands where you lose. In these situations, you'll want to play strong hole cards (high cards and big pairs), which typically start off best but whose value diminishes the further a hand progresses (due to the possibility of speculative hands improving to beat them). Be happy if your opponents continue to play speculative hands with small stack sizes – they won't make a big hand frequently enough to cover their losses.

Putting all this together, we see that in our hand example we have: deep stacks (100BB+; speculative hands are viable), excellent position (dealer; again increasing our playable hole card range), and an intermediate number of players (six; neutral). We don't yet know much about the other players' playing styles except for the EP, who is 'good' (and we'll take this to mean that he plays tight because the other players are likely fish and therefore play loose). In this situation, I'd play around 1/3 of the possible starting hands as the dealer, about 20% in EP and play the blinds based on pot odds (discussed below).

Before we continue with our sample hand, let's see some examples of actual hand ranges.

Preflop Hand Ranges

The following list gives you some example hand ranges that you can use as a starting point, both to determine which hands are playable in a specific situation, and also to put your opponents on a range. Note that you could make a range of a given size that is made up of different card combinations than those given here. I'm giving typical combinations.

Cards are abbreviated by the numbers 2-9, T for ten, J for Jack, Q for Queen, K for King and A for Ace. Hole cards of the same suit are suited and are indicated with an 's', e.g., AKs; hole cards of differing suits are offsuit and are indicated with an 'o', e.g., AKo. A '+' means the indicated strength or better, e.g. 88+ means a pair of eights or a higher pair. Broadway means a card that's a ten or higher, so broadway hole cards would be JT+. An 'x' refers to any card, e.g., Ax means an ace with any other card.

10% . . . 88+, all suited aces, AQo+, KQs, QJs

15% . . . 66+, all suited aces, ATo+, KJs+, QJs, KJo+, QJo, suited connectors T9s+

20% . . . all pairs, all suited aces, all broadways, suited connectors 87s+, suited gappers 97s+

25% . . . all pairs, all suited aces, all broadways, T9o, suited connectors 43s+, suited gappers 64s+

30% . . . all pairs, all suited aces, K8s+, A8o+, all broadways, T9o, suited connectors 43s+, suited gappers 53s+, suited double-gappers 74s+

35% . . . all pairs, all suited aces, all suited kings, A7o+, A5o, all broadways, suited connectors 43s+, suited gappers 53s+, suited double-gappers 63s+, non-suited connectors 87o+

40% . . . all pairs, all aces, all suited kings, all broadways, all suited connectors, all suited gappers, suited double-gappers 63s+, non-suited connectors 76o+

50% . . . all pairs, all suited cards, all Ax, all broadways, non-suited connectors 65o+

Going back to our example hand, we already know that the good player opened (raised the big blind) from first position, so given that he's playing tight, we'll put him on a range of around the best 20% of hole cards. From the list above, we see that this likely includes all pairs, all suited aces, all broadways, suited connectors 87s+ and some high suited-gappers. Against this range of hands, our AA have around an 84% chance to win if the hand goes to showdown (the river). The weak player in the cutoff position who just called likely has a much larger range of hands, but as he didn't raise he probably has nothing spectacular (although many casual players won't re-raise without a hand as strong as AA or KK). We'll put him on all but the best and worst hands, although assuming any two cards wouldn't be far off either. Against both these players combined, we're about a 2-1 favourite to win.

I've already mentioned pot odds, and now is a good time to discuss this concept in greater detail. It's quite simple, but still important. Let's say we decide to bet on the toss of a fair coin. If you bet $1 on heads, and I bet $1 on tails, we each have a 50% chance of winning $1 and a 50% chance of losing $1. This means that, over a long series of coin flips, we can expect to break even. Financially, there's no reason to take this bet, and no reason not to. But if we keep the rules the same, but this time I put up $1.10 and you still put up $1, now you should take the bet (as many times as I'll agree to it). Over time, you'll make a 5% profit. Note: in practice, you could still end up losing money over the short run, but the more you play, the closer to that 5% expected profit you'll get.

Now let's say we roll a die. You put up $1, and if you roll a 6 I give you $5, otherwise you lose the dollar. Should you take the bet? 5/6 times you lose $1, the other 1/6 you win $5 (plus your $1 stake which you also get back). Again this is a break-even proposition. We say that I'm offering you odds of five to one (5-1) against rolling a six, since for every $1 you bet, you win $5 when you roll a six. Note that 5-1 odds is equivalent to a 1 in 6 chance. Don't confuse the two: 1/6 means that something happens 1 time out of every six chances, which is the same as saying it happens 1 time and doesn't happen 5 times, hence 5-1 odds against it happening.

In order to be on the winning side when gambling, you need to have the odds in your favour. This is how casinos in Las Vegas make their money. For instance, in roulette you can bet on number 13 and you'll be given odds of 35-1. However, in addition to the numbers 1-36, the roulette wheel also has the numbers 0 and 00. So to be fair the casino should offer you 37-1, but as they want to make money they don't do this. Similarly, if you bet on red, you're offered even money (1-1), but because the 0 and 00 are green, 18 numbers are black and the other 18 numbers are red, you only have an 18/38 = 47.37% chance to win. This modest house edge multiplied by the huge volume of bets placed by the casino's customers makes billions of dollars of profit for casinos every year. Other casino games also have a small, built-in edge to the house, with similar results.

Poker is different, though, because in addition to chance there is a large element of skill involved. The players who make better decisions will, over time, take money off the players who make worse decisions. It should now be clear that a good decision in this context means betting when the odds are in your favour. Being a good poker player is like being the casino in roulette: you have a built-in edge that will make you money.

Let's take an easy example. We're playing NLHE and we have 9h8h. The board is Ah Kc 2h 5s. We currently have a flush draw, with just the river card to come. There's one opponent left in the hand, who bets his last $1 into a pot of $3. From his actions during the hand, we think he's got a pair of aces. Should we call (remember, we can't raise since he's all-in)? First, note that we're getting odds of 4-1 (the pot is $4 and we must put in $1 to call). It's a safe assumption that if we don't hit the flush we'll lose the hand, and we'll assume that if we do hit the flush we'll win. So, to see if these pot odds are sufficient to justify a call, we need to know the number of cards left in the deck that will give us the flush (which poker players refer to as the number of outs). A pack of cards contains 13 cards of each of the four suits. Four hearts are out already (two on the board and our two hole cards). That leaves 9 hearts in the deck (9 outs). There are 46 out of 52 cards remaining in the deck, so that gives us a 9/46 chance = 37-9 odds against hitting our flush. This is slightly over 4-1 (about 4.1-1), so in fact we shouldn't call in this case, since we need the odds to be below the pot odds of 4-1. If our opponent had gone all-in with $2 into a pot of $4, he'd only be offering us odds of 6-2 (= 3-1), and the fold would be even clearer. Conversely, if he'd only had 50¢ left, and bet that into $4, we'd be getting 9-1 and it would be an easy call. So it's vital to always be aware of the pot odds when making decisions during each hand you play.

This naturally leads us to conclude that, in no-limit poker, we can often choose a bet size that doesn't give our opponent the correct odds to call – we can induce him to make a mistake. And his mistakes are our profit. This is one of the ways we went wrong in the first run-through of our sample hand. On the turn with our pair of aces, EP had checked to us and we only bet $1 into a pot of $4.05, thus offering our opponent odds greater than 5-1 to hit his straight. Even though it was a 5-1 shot against him getting a straight on the river, he knew that there was a real chance that we'd call a river bet if he got the straight, so the implied odds (which are the pot odds plus any extra value we get from bets when we make a winning hand) were better than 5-1. His assessment proved to be accurate, as he got us to call another $3.50 after he made his straight on the river.

So with all this insight, the correct pre-flop play in our sample hand where we have the two aces is clear: we want to make a healthy re-raise. Ideally, the first player would re-raise back and all the money would go in the pot pre-flop. If this happened against a good player who was deep stacked, it's extremely likely that you've got him crushed (he'd most likely have KK, but QQ and AK are also possible; all these possibilities play terribly against AA. Of course he could also have AA, but this is extremely unlikely as between us we'd have to have all the aces). So we raise. But how much to raise? To know this, we need to know what pot odds to offer him that would be favourable to us; and to know this, we must be aware of several pieces of information.

Firstly, what is the effective stack size? This is equal to the smallest stack of chips that any of the players in the hand has in front of them (and therefore the maximum that can be bet on the hand). For the sake of argument, we'll say that all live players in this hand have $10. Next we need to know the pot size. There is 15¢ from the blinds, 40¢ from the raiser and another 40¢ from the caller, which totals 95¢. This gives us a baseline to work from (because we'll usually bet a certain fraction of the pot size, depending on the pot odds we want to offer). All the players are still deep stacked (with a pot-to-stack ratio above 10).

Next we need to know how many other players are in the hand. There is the raiser and the caller of course, but don't forget that both blinds are still to play. If we just call here, the pot size would go up to $1.35, and the small blind would have to put another 35¢ in to call, so he'd be getting odds of 135-35 or almost 4-1. This would mean he should call with a fairly wide range of hole cards. And if the small blind called, the big blind would have to put 30¢ into a pot of $1.70, which offers him almost 6-1, added to which he is the last to speak so he knows that he'd be closing the betting round (the preflop betting would be concluded and the flop would be dealt). He'd call with just about any two cards. As we established above, having this many opponents would be awful for us with our two aces. We'd like the blinds to go away, and we'd like to get some value off one of the other two players. Ideally we'd like one or two callers, rather than having everyone fold. As we're in a home game where players tend to make loose calls (they play multiple streets on many hands), I'd go with a fairly large bet here, around pot-size or even slightly higher. High enough to offer poor odds to speculative hands, but not so high that we scare everyone off. So we re-raise $1.

As in the first example, the small blind folds. But this time the big blind also folds. In the previous example, the BB had to call 30¢ into a pot of $1.35. He was getting odds of 9-2 and he was closing the action, so he could reasonably do this with many starting hands. But in this example, he's seen an open-raise followed by a call then a re-raise from the dealer. Now he'd have to call 90¢ in a pot of $1.95, only getting odds a little over 2-1. But even worse, there are three players left to speak behind him, two of whom could re-raise again, and two of whom have already raised (suggesting they have strong hole cards). Now the BB will fold a large proportion of the hands he'd have called with previously. And indeed he correctly elects to fold (his Ad5d) here.

Next to speak is the EP, who made the original open-raise. This is more interesting. We know from before that he's got QhJh, and he's also out of position against us (he must act before us on each betting round), and we've shown a lot of strength by re-raising his open-raise (he knows we play tight and thus we'd rarely re-raise him with a weak hand as a bluff. The two of us are the best players at the table, so there's little to be gained by playing fast against each other with marginal hands when the other players are easy to beat). His QhJh is objectively quite weak in this situation. Even if he was to make top pair with a queen on the flop, he'd have little confidence that his hand was the strongest, because a lot of our hole cards would have his cards dominated (i.e., we could hit the same pair but with a higher other hole card (kicker)). In addition to having an overpair with AA or KK, we could have top pair with a better kicker (AQ or KQ) or a set with QQ (and similar considerations apply if it was a J-high flop; our range certainly includes AJ, KJ and JJ), so on a Q- or J-high flop he'd often still be second best. In fact, against our likely range of cards (i.e. the hole cards we'd put in a re-raise with), he's got about a 1/3 chance of beating us. Realistically, he'd want two pairs or better on the flop, or at least a flush draw or an open-ended straight draw. He's also still got to worry about the cutoff, who is still live in the hand. So calling is marginal at best. It would be very reckless of him to put in another re-raise here (unless we've shown a tendency to regularly re-raise on a bluff). In an actual situation like this, a good player would tend to fold.

EP indeed folds, the cutoff also folds, and we take a pot of 95¢. Unfortunately, we didn't get any callers, but we still took down a small pot, and avoided exposing ourselves to losing a big one by letting our opponents draw cheaply. Contrast this result with what happened in the first example, when we played the hand passively (we lost $6.10).

At this point, let's think more about the differences between individual players. In our first example we considered player types in broad terms, but of course every player is different and thus they handle specific situations differently. Every player projects a certain image to the other players (tight, loose, weak, strong, predictable, wild, etc.). We must include ourself in this; the other players will think of us as a certain type of player, based on our personality and actions at the table. A player with a reputation for bluffing too much will get their bets and raises called far more often than a player who is usually very timid. However, note that many casual players are more or less oblivious both to the image they project to others and to the images of their opponents. That's why, for a small stakes home game, we can just put our opponents into one of a few simple categories and play them accordingly. Just bear in mind that some players actually know what they're doing, and if one of these players is in your game, you must be aware of your image when you play a hand with them. Furthermore, if you closely observe your opponents, you'll probably pick up on specific weaknesses in the way they play and be able to exploit them. For example, a common weakness is for a player to