Inferring Evolutionary Relationships from Sequences

Describe asks for the characteristics and features of the method. It must be specific enough that another student could follow it, and it must say what is done with the numbers, not only how they are obtained.

5 marks. State why the same protein is used, then the alignment and counting, then the matrix, then the tree, then the reading rule.

Use the same protein from all four species, and a protein present across a wide range of organisms because it performs an essential function, so that the sequences are homologous, meaning they descend from one ancestral gene. Comparing different proteins, or unrelated sequences, would produce differences that reflect the proteins rather than the species. A highly conserved protein is preferred because most of its sequence is constrained by function, so the differences that do accumulate are informative rather than random noise.

Write the four sequences one beneath another and align them position by position, inserting gaps where necessary so that corresponding positions line up. Then compare each pair of species and count the number of positions at which their amino acids differ. This gives a difference count for every possible pair.

Record the counts in a difference matrix, a table with the species along both axes and the number of differences in each cell. This makes the pattern visible: the smallest counts identify the most similar pairs and the largest counts the most divergent.

Interpret the counts using the principle that differences accumulate over time since two lineages diverged: the fewer the differences between two species, the more recently they shared a common ancestor. Rank the pairs from fewest differences to most, and build a tree by joining the most similar pair first, then attaching the next most similar species to that group, and so on, so that the species most different from all the others joins the tree last and at its base.

Finally, check the tree reproduces the matrix: species grouped closely together should have small counts between them, and a species placed at the base should have large counts against every other. The method assumes that differences accumulate at a broadly similar rate across the lineages compared, and that the positions counted are changing independently rather than being constrained together, so the inferred tree should be treated as a hypothesis to be tested against other proteins and against anatomical evidence.

Deduce asks you to reach a conclusion by reasoning from given information. The reasoning must be shown, and because the question asks for assumptions, an answer without them is incomplete however good the arithmetic.

5 marks. Establish the rate from the calibration point, apply it to each unknown, state the answers with appropriate precision, check they are plausible, then state the assumptions.

W and X differ at 3 positions and are known from fossil evidence to have diverged 15 million years ago. The rate of accumulation of differences is therefore 3 differences divided by 15 million years, which is 0.2 differences per million years.

It matters that the rate was calibrated from a pairwise count and is being applied to pairwise counts. Differences accumulate independently along both lineages after they diverge, so the number of differences between two living species reflects change on two branches, not one, and is proportional to roughly twice the time since divergence. Because the calibration was made the same way, that factor is already inside the 0.2 figure and cancels out. Taking a per-lineage mutation rate from elsewhere and applying it to a pairwise count would overestimate every divergence time by about a factor of two, and nothing in the arithmetic would reveal the error.

W and Y differ at 18 positions. Dividing 18 differences by the rate of 0.2 differences per million years gives 18 divided by 0.2, which is 90 million years. The divergence of Y from W is therefore estimated at approximately 90 million years ago.

W and Z differ at 41 positions. Dividing 41 by 0.2 gives 205 million years, so the divergence of Z from W is estimated at approximately 205 million years ago. The order of relatedness is therefore X most closely related to W, then Y, then Z most distantly, which matches the order of the difference counts as it must.

The figures should be checked for plausibility before being reported. Divergences of 90 and 205 million years are within the range spanned by vertebrate groups, so both are possible. Two results would not be: a divergence time greater than the age of life on Earth, or one exceeding the age of the group the species belong to, and a divergence estimate older than the oldest fossil of the whole lineage would indicate an error in the rate or the counts rather than a discovery. Note also that 41 differences in a 100-position protein means the sequences are 59 per cent identical, so the count is well short of the point at which so many positions have changed that the relationship is no longer readable.

The deduction assumes that differences have accumulated at an approximately constant rate over the whole period and in all four lineages, which is the molecular clock assumption and is only approximately true, since rates vary between proteins, between lineages and over time; that the fossil calibration of 15 million years is itself accurate, so any error in it scales all the derived dates; that each counted position has changed only once, which becomes unsafe as counts rise; and that the sequences are homologous and correctly aligned. Because of these, the figures should be reported as estimates of order of magnitude rather than as precise dates, and they would be far more secure if the same calculation were repeated using several independent proteins.