Alignment-Integrated Reconstruction of Ancestral Sequences Improves Accuracy

Ancestral sequence reconstruction (ASR) uses an alignment of extant protein sequences, a phylogeny describing the history of the protein family and a model of the molecular-evolutionary process to infer the sequences of ancient proteins, allowing researchers to directly investigate the impact of sequence evolution on protein structure and function. Like all statistical inferences, ASR can be sensitive to violations of its underlying assumptions. Previous studies have shown that, whereas phylogenetic uncertainty has only a very weak impact on ASR accuracy, uncertainty in the protein sequence alignment can more strongly affect inferred ancestral sequences. Here, we show that errors in sequence alignment can produce errors in ASR across a range of realistic and simplified evolutionary scenarios. Importantly, sequence reconstruction errors can lead to errors in estimates of structural and functional properties of ancestral proteins, potentially undermining the reliability of analyses relying on ASR. We introduce an alignment-integrated ASR approach that combines information from many different sequence alignments. We show that integrating alignment uncertainty improves ASR accuracy and the accuracy of downstream structural and functional inferences, often performing as well as highly accurate structure-guided alignment. Given the growing evidence that sequence alignment errors can impact the reliability of ASR studies, we recommend that future studies incorporate approaches to mitigate the impact of alignment uncertainty. Probabilistic modeling of insertion and deletion events has the potential to radically improve ASR accuracy when the model reflects the true underlying evolutionary history, but further studies are required to thoroughly evaluate the reliability of these approaches under realistic conditions.

Alignment-Integrated Reconstruction of Ancestral Sequences Improves Accuracy

Ancestral sequence reconstruction (ASR) uses an alignment of extant protein sequences, a phylogeny describing the history of the protein family and a model of the molecular-evolutionary process to infer the sequences of ancient proteins, allowing researchers to directly investigate the impact of sequence evolution on protein structure and function. Like all statistical inferences, ASR can be sensitive to violations of its underlying assumptions. Previous studies have shown that, while phylogenetic uncertainty has only a very weak impact on ASR accuracy, uncertainty in the protein sequence alignment can more strongly affect inferred ancestral sequences. Here we show that errors in sequence alignment can produce errors in ASR across a range of realistic and simplified evolutionary scenarios. Importantly, sequence reconstruction errors can lead to errors in estimates of structural and functional properties of ancestral proteins, potentially undermining the reliability of analyses relying on ASR. We introduce an alignment-integrated ASR approach that combines information from many different sequence alignments. We show that integrating alignment uncertainty improves ASR accuracy and the accuracy of downstream structural and functional inferences, often performing as well as highly-accurate structure-guided alignment. Given the growing evidence that sequence alignment errors can impact the reliability of ASR studies, we recommend that future studies incorporate approaches to mitigate the impact of alignment uncertainty. Probabilistic modeling of insertion and deletion events has the potential to radically improve ASR accuracy when the model reflects the true underlying evolutionary history, but further studies are required to thoroughly evaluate the reliability of these approaches under realistic conditions.

The effect of alignment uncertainty, substitution models and priors in building and dating the mammal tree of life

Background The flood of genomic data to help build and date the tree of life requires automation at several critical junctures, most importantly during sequence assembly and alignment. It is widely appreciated that automated alignment protocols can yield inaccuracies, but the relative impact of various sources error on phylogenomic analysis is not yet known. This study employs an updated mammal data set of 5162 coding loci sampled from 90 species to evaluate the effects of alignment uncertainty, substitution models, and fossil priors on gene tree, species tree, and divergence time estimation. Additionally, a novel coalescent likelihood ratio test is introduced for comparing competing species trees against a given set of gene trees. Results The aligned DNA sequences of 5162 loci from 90 species were trimmed and filtered using trimAL and two filtering protocols. The final dataset contains 4 sets of alignments - before trimming, after trimming, filtered by a recently proposed pipeline, and further filtered by comparing ML gene trees for each locus with the concatenation tree. Our analyses suggest that the average discordance among the coalescent trees is significantly smaller than that among the concatenation trees estimated from the 4 sets of alignments or with different substitution models. There is no significant difference among the divergence times estimated with different substitution models. However, the divergence dates estimated from the alignments after trimming are more recent than those estimated from the alignments before trimming. Conclusions Our results highlight that alignment uncertainty of the updated mammal data set and the choice of substitution models have little impact on tree topologies yielded by coalescent methods for species tree estimation, whereas they are more influential on the trees made by concatenation. Given the choice of calibration scheme and clock models, divergence time estimates are robust to the choice of substitution models, but removing alignments deemed problematic by trimming algorithms can lead to more recent dates. Although the fossil prior is important in divergence time estimation, Bayesian estimates of divergence times in this data set are driven primarily by the sequence data.