Off-topic here, but this story is really striking and has a profound impact.
A mini-review. Coming out soon. I am still trying to figure out if it's ok to put a preprint online here, considering that it might take another 1-2 months for the journal to publish it online. Anyway, below is the abstract:
Recently, an interesting question has emerged in the evolutionary interpretation of sequence substitution data as evidence of amino acid selection pressure. Specifically, the Ka/Ks metric was designed to measure selection pressure on amino acid substitutions, assuming that the synonymous substitution rate Ks reflects the neutral nucleotide substitution rate. However, there is increasing evidence for selection pressure at silent sites due to constraints of RNA splicing. Is Ka/Ks an appropriate metric for selection pressure on amino acid substitutions, in the presence of other selection pressures acting only at the RNA level (such as selection for exonic splicing enhancers)? Or can the resulting decreases in Ks from such selection pressures introduce bias into the Ka/Ks metric, so that it no longer gives an accurate measure of amino acid level selection pressure? In this review we present both mathematical models and empirical evidence for these divergent points of view.
Here is the reason why this is an important question:
Recently, many different types of studies have reported widespread selection pressure against synonymous substitutions in the mammalian genomes, due to constraints of pre-mRNA splicing (e.g. Cartegni et al., 2002; Buratti and Baralle, 2004; Fairbrother et al., 2004; Pagani, 2004; Baek and Green, 2005; Carlini and Genut, 2005; Ohler et al., 2005; Pagani et al., 2005; Parmley et al., 2005; Xing and Lee, 2005).Thus, the standard assumption of Ks as a measure of selectively neutral background mutation has become questionable and in many cases demonstrably false. Since this assumption is the basis of the well-known Ka/Ks metric of amino acid selection pressure, this is very problematic, especially given the increasingly ubiquitous use of Ka/Ks to study gene function. Unfortunately, these considerations have not been integrated into current thinking about Ka/Ks, and have produced some serious questions in the literature. For example, Pagani and colleagues commented in their article (Synonymous mutations in CFTR exon 12 affect splicing and are not neutral in evolution, PNAS, 2005, May 2, 102(18):6368-72.): "...because our results show that synonymous mutations are not strictly neutral but constrained by splicing requirement, the Ka/Ks ratios may yield inaccurate estimates of evolutionary rates...the presumed Ks evolutionary neutrality should be reevaluated in the light of splicing requirements." Similarly, Baek and Green wrote (PNAS, 2005 Sep 6;102(36):12813-8.): "...One implication is that the common practice of using Ks to represent the neutral rate "baseline" for the purpose of assessing selection on protein coding sequences with the Ka/Ks ratio can be misleading...". Apart from such ambivalent statements of concern, there has been no explicit analysis or review of the detailed impact of RNA selection pressure on the Ka/Ks metric. This question of the impact of RNA selection pressure on Ka/Ks must be raised explicitly, in a review that can clarify both the problem and what must be done to answer it.
I am going to San Diego tonight for the Recomb workshop on systems biology and regulatory genomics. Featuring
an excellent list of talks.