Limits...
Kr/Kc but not dN/dS correlates positively with body mass in birds, raising implications for inferring lineage-specific selection.

Weber CC, Nabholz B, Romiguier J, Ellegren H - Genome Biol. (2014)

Bottom Line: It is often suggested that, all else being equal, dN/dS should be lower in populations with large effective size (Ne) due to increased efficacy of purifying selection.Our results in principle accord with the notion that non-synonymous substitutions causing radical amino acid changes are more efficiently removed by selection in large populations, consistent with nearly neutral theory.These findings have implications for the use of dN/dS and suggest that caution is warranted when drawing conclusions about lineage-specific modes of protein evolution using this metric.

View Article: PubMed Central - PubMed

ABSTRACT

Background: The ratio of the rates of non-synonymous and synonymous substitution (dN/dS) is commonly used to estimate selection in coding sequences. It is often suggested that, all else being equal, dN/dS should be lower in populations with large effective size (Ne) due to increased efficacy of purifying selection. As Ne is difficult to measure directly, life history traits such as body mass, which is typically negatively associated with population size, have commonly been used as proxies in empirical tests of this hypothesis. However, evidence of whether the expected positive correlation between body mass and dN/dS is consistently observed is conflicting.

Results: Employing whole genome sequence data from 48 avian species, we assess the relationship between rates of molecular evolution and life history in birds. We find a negative correlation between dN/dS and body mass, contrary to nearly neutral expectation. This raises the question whether the correlation might be a method artefact. We therefore in turn consider non-stationary base composition, divergence time and saturation as possible explanations, but find no clear patterns. However, in striking contrast to dN/dS, the ratio of radical to conservative amino acid substitutions (Kr/Kc) correlates positively with body mass.

Conclusions: Our results in principle accord with the notion that non-synonymous substitutions causing radical amino acid changes are more efficiently removed by selection in large populations, consistent with nearly neutral theory. These findings have implications for the use of dN/dS and suggest that caution is warranted when drawing conclusions about lineage-specific modes of protein evolution using this metric.

Show MeSH

Related in: MedlinePlus

Saturation for substitution at third sites and amino acid changes. Uncorrected observed differences are smaller than estimated phylogenetic distances for third codon sites and amino acids. The red line denotes the relationship expected in the absence of saturation (x=y), as the phylogenetic and observed distances would be equal in this case. Divergence for third codon positions is greater than for amino acids, particularly in orthologues with high variance in GC3 between species.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4264323&req=5

Fig5: Saturation for substitution at third sites and amino acid changes. Uncorrected observed differences are smaller than estimated phylogenetic distances for third codon sites and amino acids. The red line denotes the relationship expected in the absence of saturation (x=y), as the phylogenetic and observed distances would be equal in this case. Divergence for third codon positions is greater than for amino acids, particularly in orthologues with high variance in GC3 between species.

Mentions: To assess whether there is evidence for saturation in our data, we compared the phylogenetic distance (the sum of branch lengths between two given species) to the number of uncorrected pairwise differences for high- and low-variance sequences, as considering the full data set would not have been computationally tractable. That the uncorrected distance does not increase linearly with the corrected distance for the high-variance subset, instead remaining lower (Figure 5), indicates that there are multiple hits. As expected, divergence for third codon positions is greater than for amino acids. This implies that a degree of saturation and therefore underestimation of dS relative to dN might be of concern for our data. The weaker signal of saturation at third positions relative to amino acids in the lower-variance subset is consistent with the shorter branch lengths observed here (Figure 5). Constraint cannot explain the patterns we observe in the saturation plots, as it would affect both observed and phylogenetic distances. Nevertheless, the extent to which saturation affects our estimates of dN/dS is not clear.Figure 5


Kr/Kc but not dN/dS correlates positively with body mass in birds, raising implications for inferring lineage-specific selection.

Weber CC, Nabholz B, Romiguier J, Ellegren H - Genome Biol. (2014)

Saturation for substitution at third sites and amino acid changes. Uncorrected observed differences are smaller than estimated phylogenetic distances for third codon sites and amino acids. The red line denotes the relationship expected in the absence of saturation (x=y), as the phylogenetic and observed distances would be equal in this case. Divergence for third codon positions is greater than for amino acids, particularly in orthologues with high variance in GC3 between species.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4264323&req=5

Fig5: Saturation for substitution at third sites and amino acid changes. Uncorrected observed differences are smaller than estimated phylogenetic distances for third codon sites and amino acids. The red line denotes the relationship expected in the absence of saturation (x=y), as the phylogenetic and observed distances would be equal in this case. Divergence for third codon positions is greater than for amino acids, particularly in orthologues with high variance in GC3 between species.
Mentions: To assess whether there is evidence for saturation in our data, we compared the phylogenetic distance (the sum of branch lengths between two given species) to the number of uncorrected pairwise differences for high- and low-variance sequences, as considering the full data set would not have been computationally tractable. That the uncorrected distance does not increase linearly with the corrected distance for the high-variance subset, instead remaining lower (Figure 5), indicates that there are multiple hits. As expected, divergence for third codon positions is greater than for amino acids. This implies that a degree of saturation and therefore underestimation of dS relative to dN might be of concern for our data. The weaker signal of saturation at third positions relative to amino acids in the lower-variance subset is consistent with the shorter branch lengths observed here (Figure 5). Constraint cannot explain the patterns we observe in the saturation plots, as it would affect both observed and phylogenetic distances. Nevertheless, the extent to which saturation affects our estimates of dN/dS is not clear.Figure 5

Bottom Line: It is often suggested that, all else being equal, dN/dS should be lower in populations with large effective size (Ne) due to increased efficacy of purifying selection.Our results in principle accord with the notion that non-synonymous substitutions causing radical amino acid changes are more efficiently removed by selection in large populations, consistent with nearly neutral theory.These findings have implications for the use of dN/dS and suggest that caution is warranted when drawing conclusions about lineage-specific modes of protein evolution using this metric.

View Article: PubMed Central - PubMed

ABSTRACT

Background: The ratio of the rates of non-synonymous and synonymous substitution (dN/dS) is commonly used to estimate selection in coding sequences. It is often suggested that, all else being equal, dN/dS should be lower in populations with large effective size (Ne) due to increased efficacy of purifying selection. As Ne is difficult to measure directly, life history traits such as body mass, which is typically negatively associated with population size, have commonly been used as proxies in empirical tests of this hypothesis. However, evidence of whether the expected positive correlation between body mass and dN/dS is consistently observed is conflicting.

Results: Employing whole genome sequence data from 48 avian species, we assess the relationship between rates of molecular evolution and life history in birds. We find a negative correlation between dN/dS and body mass, contrary to nearly neutral expectation. This raises the question whether the correlation might be a method artefact. We therefore in turn consider non-stationary base composition, divergence time and saturation as possible explanations, but find no clear patterns. However, in striking contrast to dN/dS, the ratio of radical to conservative amino acid substitutions (Kr/Kc) correlates positively with body mass.

Conclusions: Our results in principle accord with the notion that non-synonymous substitutions causing radical amino acid changes are more efficiently removed by selection in large populations, consistent with nearly neutral theory. These findings have implications for the use of dN/dS and suggest that caution is warranted when drawing conclusions about lineage-specific modes of protein evolution using this metric.

Show MeSH
Related in: MedlinePlus