Viral genome sequence datasets display pervasive evidence of strandspecific substitution biases that are best described using nonreversible nucleotide substitution models
Curation statements for this article:
Curated by eLife
eLife assessment
This valuable study revisits the effects of substitution model selection on phylogenetics by comparing reversible and nonreversible DNA substitution models. The authors provide evidence that 1) non timereversible models sometimes perform better than general timereversible models when inferring phylogenetic trees out of simulated viral genome sequence data sets, and that 2) non timereversible models can fit the real data better than the reversible substitution models commonly used in phylogenetics, a finding consistent with previous work. However, the methods are incomplete in supporting the main conclusion of the manuscript, that is that non timereversible models should be incorporated in the model selection process for these data sets.
This article has been Reviewed by the following groups
Listed in
 Evaluated articles (eLife)
Abstract
The vast majority of phylogenetic trees are inferred from molecular sequence data (nucleotides or amino acids) using timereversible evolutionary models which assume that, for any pair of nucleotide or amino acid characters, the relative rate of X to Y substitution is the same as the relative rate of Y to X substitution. However, this reversibility assumption is unlikely to accurately reflect the actual underlying biochemical and/or evolutionary processes that lead to the fixation of substitutions. Here, we use empirical viral genome sequence data to reveal that evolutionary nonreversibility is pervasive among most groups of viruses. Specifically, we consider two nonreversible nucleotide substitution models: (1) a 6rate nonreversible model (NREV6) in which WatsonCrick complementary substitutions occur at identical relative rates and which might therefor be most applicable to analyzing the evolution of genomes where both complementary strands are subject to the same mutational processes (such as might be expected for doublestranded (ds) RNA or dsDNA genomes); and (2) a 12rate nonreversible model (NREV12) in which all relative substitution types are free to occur at different rates and which might therefore be applicable to analyzing the evolution of genomes where the complementary genome strands are subject to different mutational processes (such as might be expected for viruses with singlestranded (ss) RNA or ssDNA genomes).Using likelihood ratio and Akaike Information Criterionbased model tests, we show that, surprisingly, NREV12 provided a significantly better fit to 21/31 dsRNA and 20/30 dsDNA datasets than did the general time reversible (GTR) and NREV6 models with NREV6 providing a better fit than NREV12 and GTR in only 5/30 dsDNA and 2/31 dsRNA datasets. As expected, NREV12 provided a significantly better fit to 24/33 ssDNA and 40/47 ssRNA datasets. Next, we used simulations to show that increasing degrees of strandspecific substitution bias decrease the accuracy of phylogenetic inference irrespective of whether GTR or NREV12 is used to describe mutational processes. However, in cases where strandspecific substitution biases are extreme (such as in SARSCoV2 and Torque teno sus virus datasets) NREV12 tends to yield more accurate phylogenetic trees than those obtained using GTR.We show that NREV12 should, be seriously considered during the model selection phase of phylogenetic analyses involving viral genomic sequences.
Article activity feed



eLife assessment
This valuable study revisits the effects of substitution model selection on phylogenetics by comparing reversible and nonreversible DNA substitution models. The authors provide evidence that 1) non timereversible models sometimes perform better than general timereversible models when inferring phylogenetic trees out of simulated viral genome sequence data sets, and that 2) non timereversible models can fit the real data better than the reversible substitution models commonly used in phylogenetics, a finding consistent with previous work. However, the methods are incomplete in supporting the main conclusion of the manuscript, that is that non timereversible models should be incorporated in the model selection process for these data sets.

Reviewer #1 (Public Review):
The study by SiangaMete et al revisits the effects of substitution model selection on phylogenetics by comparing reversible and nonreversible DNA substitution models. This topic is not new, previous works already showed that nonreversible, and also covarion, substitution models can fit the real data better than the reversible substitution models commonly used in phylogenetics. In this regard, the results of the present study are not surprising. Specific comments are shown below.
Major comments
It is well known that nonreversible models can fit the real data better than the commonly used reversible substitution models, see for example,
https://academic.oup.com/sysbio/article/71/5/1110/6525257
https://onlinelibrary.wiley.com/doi/10.1111/jeb.14147?af=R
The manuscript indicates that the results (better …Reviewer #1 (Public Review):
The study by SiangaMete et al revisits the effects of substitution model selection on phylogenetics by comparing reversible and nonreversible DNA substitution models. This topic is not new, previous works already showed that nonreversible, and also covarion, substitution models can fit the real data better than the reversible substitution models commonly used in phylogenetics. In this regard, the results of the present study are not surprising. Specific comments are shown below.
Major comments
It is well known that nonreversible models can fit the real data better than the commonly used reversible substitution models, see for example,
https://academic.oup.com/sysbio/article/71/5/1110/6525257
https://onlinelibrary.wiley.com/doi/10.1111/jeb.14147?af=R
The manuscript indicates that the results (better fitting of nonreversible models compared to reversible models) are surprising but I do not think so, I think the results would be surprising if the reversible models provide a better fitting.
I think the introduction of the manuscript should be increased with more information about nonreversible models and the diverse previous studies that already evaluated them. Also I think the manuscript should indicate that the results are not surprising, or more clearly justify why they are surprising.In the introduction and/or discussion I missed a discussion about the recent works on the influence of substitution model selection on phylogenetic tree reconstruction. Some works indicated that substitution model selection is not necessary for phylogenetic tree reconstruction,
https://academic.oup.com/mbe/article/37/7/2110/5810088
https://www.nature.com/articles/s4146701908822w
https://academic.oup.com/mbe/article/35/9/2307/5040133
While others indicated that substitution model selection is recommended for phylogenetic tree reconstruction,
https://www.sciencedirect.com/science/article/pii/S0378111923001774
https://academic.oup.com/sysbio/article/53/2/278/1690801
https://academic.oup.com/mbe/article/33/1/255/2579471
The results of the present study seem to support this second view. I think this study could be improved by providing a discussion about this aspect, including the specific contribution of this study to that.The real data was downloaded from Los Alamos HIV database. I am wondering if there were any criterion for selecting the sequences or if just all the sequences of the database for every studied virus category were analysed. Also, was any quality filter applied? How gaps and ambiguous nucleotides were considered? Notice that these aspects could affect the fitting of the models with the data.
How the nonreversible model and the data are compared considering the nonreversible substitution process? In particular, given an input MSA, how to know if the nucleotide substitution goes from state x to state y or from state y to state x in the real data if there is not a reference (i.e., wild type) sequence? All the sequences are mutants and one may not have a reference to identify the direction of the mutation, which is required for the nonreversible model. Maybe one could consider that the most abundant state is the wild type state but that may not be the case in reality. I think this is a main problem for the practical application of nonreversible substitution models in phylogenetics.

Reviewer #2 (Public Review):
The authors evaluate whether non time reversible models fit better data presenting strandspecific substitution biases than time reversible models. Specifically, the authors consider what they call NREV6 and NREV12 as candidate non timereversible models. On the one hand, they show that AIC tends to select NREV12 more often than GTR on real virus data sets. On the other hand, they show using simulated data that NREV12 leads to inferred trees that are closer to the true generating tree when the data incorporates a certain degree of non timereversibility. Based on these two experimental results, the authors conclude that "We show that nonreversible models such as NREV12 should be evaluated during the model selection phase of phylogenetic analyses involving viral genomic sequences". This is a valuable …
Reviewer #2 (Public Review):
The authors evaluate whether non time reversible models fit better data presenting strandspecific substitution biases than time reversible models. Specifically, the authors consider what they call NREV6 and NREV12 as candidate non timereversible models. On the one hand, they show that AIC tends to select NREV12 more often than GTR on real virus data sets. On the other hand, they show using simulated data that NREV12 leads to inferred trees that are closer to the true generating tree when the data incorporates a certain degree of non timereversibility. Based on these two experimental results, the authors conclude that "We show that nonreversible models such as NREV12 should be evaluated during the model selection phase of phylogenetic analyses involving viral genomic sequences". This is a valuable finding, and I agree that this is potentially good practice. However, I miss an experiment that links the two findings to support the conclusion: in particular, an experiment that solves the following question: does the bestfit model also lead to better tree topologies?
On simulated data, the significance of the difference between GTR and NREV12 inferences is evaluated using a paired t test. I miss a rationale or a reference to support that a paired t test is suitable to measure the significance of the differences of the wRF distance. Also, the results show that on average NREV12 performs better than GTR, but a pairwise comparison would be more informative: for how many sequence alignments does NREV12 perform better than GTR?
