Question

Up till now, we have assumed that all codons coding for the same amino acid are interchangeable. In fact, this not entirely true. In many species, it has been observed that certain codons are (slightly) favored. This is called codon bias. For example, AAA and AAG both code for lysine. If, in a particular species, the AAA codon is used preferentially, we would call that codon bias. One source of codon bias has to do with the efficiency of the transcriptional machinery. It is thought that favored codons can allow the transcriptional machinery to work more smoothly. This effect however, varies between species. A. The impact of codon bias on gene finding. Consider three cases. Case 1. We take a well studied species with good gene annotations (species A) and build a Markov model on codons. We then use this model to find genes in the closely related species B. These two species have little or no codon bias. Case 2. We take a well studied species with good gene annotations (species C) and build a Markov model on codons. We then use this model to find genes in the closely related species D. Species C and D each have strong codon bias, however the biases in the two species are different. (Taking the lysine example to illustrate, in one species the AAA codon is favored, but in the other the AAG is.) Case 3. We take a well studied species with good gene annotations (species E) and build a Markov model on codons. We then use this model to find genes in the closely related species F. Species E and F have strong codon bias which is the same in both species. (e.g. in the lysine example both favor AAA). Rank these three cases in terms of how effective you would expect the gene finder to be. (Effectiveness for the gene finder means having a high true positive rate and low false discovery rate.) Explain your reasoning.

Answer 1

In my viewpoint

Case 2 will be the most effective followed by case 3 and then comes case 1 in case of gene finder. The reason being in case 2 we can find all the preffered genes for the codons by utilising two closely related species (C& D) with different preferrence over the codons. In case 2 same codons will be found in the two experimental species (E & F) and in case 3 for species A & B as there is no codon bias any of the codon can be selected for any amino acids.