Evaluating Statistical Multiple Sequence Alignment in Comparison to Other Alignment Methods on Protein Data Sets

Phylogeny, biogeography and ecological diversification of New Caledonian palms (Arecaceae)

BACKGROUND AND AIMS

The geographical origin and evolutionary mechanisms underpinning the rich and distinctive New Caledonian flora remain poorly understood. This is attributable to the complex geological past of the island and to the scarcity of well-resolved species-level phylogenies. Here, we infer phylogenetic relationships and divergence times of New Caledonian palms, which comprise 40 species. We use this framework to elucidate the biogeography of New Caledonian palm lineages and to explore how extant species might have formed.

METHODS

A phylogenetic tree including 37 New Caledonian palm species and 77 relatives from tribe Areceae was inferred from 151 nuclear genes obtained by targeted sequencing. Fossil-calibrated divergence times were estimated and ancestral ranges inferred. Ancestral and extant ecological preferences in terms of elevation, precipitation and substrate were compared between New Caledonian sister species to explore their possible roles as drivers of speciation.

KEY RESULTS

New Caledonian palms form four well-supported clades, inside which relationships are well resolved. Our results support the current classification but suggest that Veillonia and Campecarpus should be resurrected and fail to clarify whether Rhopalostylidinae is sister to or nested in Basseliniinae. New Caledonian palm lineages are derived from New Guinean and Australian ancestors, which reached the island through at least three independent dispersal events between the Eocene and Miocene. Palms then dispersed out of New Caledonia at least five times, mainly towards Pacific islands. Geographical and ecological transitions associated with speciation events differed across time and genera. Substrate transitions were more frequently associated with older events than with younger ones.

CONCLUSIONS

Neighbouring areas and a mosaic of local habitats shaped the palm flora of New Caledonia, and the island played a significant role in generating palm diversity across the Pacific region. This new spatio-temporal framework will enable population-level ecological and genetic studies to unpick the mechanisms underpinning New Caledonian palm endemism.

Structural-guided design to improve the catalytic performance of aldo-keto reductase KdAKR

Aldo-keto reductases (AKRs) are important biocatalysts that can be used to synthesize chiral pharmaceutical alcohols. In this study, the catalytic activity and stereoselectivity of a NADPH-dependent AKR from Kluyveromyces dobzhanskii (KdAKR) toward t-butyl 6-chloro (5S)-hydroxy-3-oxohexanoate ((5S)-CHOH) were improved by mutating its residues in the loop regions around the substrate-binding pocket. And the thermostability of KdAKR was improved by a consensus sequence method targeted on the flexible regions. The best mutant M6 (Y28A/L58I/I63L/G223P/Y296W/W297H) exhibited a 67-fold higher catalytic efficiency compared to the wild-type (WT) KdAKR, and improved R-selectivity toward (5S)-CHOH (dep value from 47.6% to >99.5%). Moreover, M6 exhibited a 6.3-fold increase in half-life (t1/2 ) at 40°C compared to WT. Under the optimal conditions, M6 completely converted 200 g/L (5S)-CHOH to diastereomeric pure t-butyl 6-chloro-(3R, 5S)-dihydroxyhexanoate ((3R, 5S)-CDHH) within 8.0 h, with a space-time yield of 300.7 g/L/day. Our results deepen the understandings of the structure-function relationship of AKRs, providing a certain guidance for the modification of other AKRs.

Leveraging protein language models for accurate multiple sequence alignments

Multiple sequence alignment (MSA) is a critical step in the study of protein sequence and function. Typically, MSA algorithms progressively align pairs of sequences and combine these alignments with the aid of a guide tree. These alignment algorithms use scoring systems based on substitution matrices to measure amino acid similarities. Although successful, standard methods struggle on sets of proteins with low sequence identity: the so-called twilight zone of protein alignment. For these difficult cases, another source of information is needed. Protein language models are a powerful new approach that leverages massive sequence data sets to produce high-dimensional contextual embeddings for each amino acid in a sequence. These embeddings have been shown to reflect physicochemical and higher-order structural and functional attributes of amino acids within proteins. Here, we present a novel approach to MSA, based on clustering and ordering amino acid contextual embeddings. Our method for aligning semantically consistent groups of proteins circumvents the need for many standard components of MSA algorithms, avoiding initial guide tree construction, intermediate pairwise alignments, gap penalties, and substitution matrices. The added information from contextual embeddings leads to higher accuracy alignments for structurally similar proteins with low amino-acid similarity. We anticipate that protein language models will become a fundamental component of the next generation of algorithms for generating MSAs.

Ancient Rapid Radiation Explains Most Conflicts Among Gene Trees and Well-Supported Phylogenomic Trees of Nostocalean Cyanobacteria

Prokaryotic genomes are often considered to be mosaics of genes that do not necessarily share the same evolutionary history due to widespread horizontal gene transfers (HGTs). Consequently, representing evolutionary relationships of prokaryotes as bifurcating trees has long been controversial. However, studies reporting conflicts among gene trees derived from phylogenomic data sets have shown that these conflicts can be the result of artifacts or evolutionary processes other than HGT, such as incomplete lineage sorting, low phylogenetic signal, and systematic errors due to substitution model misspecification. Here, we present the results of an extensive exploration of phylogenetic conflicts in the cyanobacterial order Nostocales, for which previous studies have inferred strongly supported conflicting relationships when using different concatenated phylogenomic data sets. We found that most of these conflicts are concentrated in deep clusters of short internodes of the Nostocales phylogeny, where the great majority of individual genes have low resolving power. We then inferred phylogenetic networks to detect HGT events while also accounting for incomplete lineage sorting. Our results indicate that most conflicts among gene trees are likely due to incomplete lineage sorting linked to an ancient rapid radiation, rather than to HGTs. Moreover, the short internodes of this radiation fit the expectations of the anomaly zone, i.e., a region of the tree parameter space where a species tree is discordant with its most likely gene tree. We demonstrated that concatenation of different sets of loci can recover up to 17 distinct and well-supported relationships within the putative anomaly zone of Nostocales, corresponding to the observed conflicts among well-supported trees based on concatenated data sets from previous studies. Our findings highlight the important role of rapid radiations as a potential cause of strongly conflicting phylogenetic relationships when using phylogenomic data sets of bacteria. We propose that polytomies may be the most appropriate phylogenetic representation of these rapid radiations that are part of anomaly zones, especially when all possible genomic markers have been considered to infer these phylogenies. [Anomaly zone; bacteria; horizontal gene transfer; incomplete lineage sorting; Nostocales; phylogenomic conflict; rapid radiation; Rhizonema.].

Highly significant improvement of protein sequence alignments with AlphaFold2

MOTIVATION

Protein sequence alignments are essential to structural, evolutionary and functional analysis, but their accuracy is often limited by sequence similarity unless molecular structures are available. Protein structures predicted at experimental grade accuracy, as achieved by AlphaFold2, could therefore have a major impact on sequence analysis.

RESULTS

Here, we find that multiple sequence alignments estimated on AlphaFold2 predictions are almost as accurate as alignments estimated on experimental structures and significantly closer to the structural reference than sequence-based alignments. We also show that AlphaFold2 structural models of relatively low quality can be used to obtain highly accurate alignments. These results suggest that, besides structure modeling, AlphaFold2 encodes higher-order dependencies that can be exploited for sequence analysis.

AVAILABILITY AND IMPLEMENTATION

All data, analyses and results are available on Zenodo (https://doi.org/10.5281/zenodo.7031286). The code and scripts have been deposited in GitHub (https://github.com/cbcrg/msa-af2-nf) and the various containers in (https://cloud.sylabs.io/library/athbaltzis/af2/alphafold, https://hub.docker.com/r/athbaltzis/pred).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Accuracy of multiple sequence alignment methods in the reconstruction of transposable element families

The construction of a high-quality multiple sequence alignment (MSA) from copies of a transposable element (TE) is a critical step in the characterization of a new TE family. Most studies of MSA accuracy have been conducted on protein or RNA sequence families, where structural features and strong signals of selection may assist with alignment. Less attention has been given to the quality of sequence alignments involving neutrally evolving DNA sequences such as those resulting from TE replication. Transposable element sequences are challenging to align due to their wide divergence ranges, fragmentation, and predominantly-neutral mutation patterns. To gain insight into the effects of these properties on MSA accuracy, we developed a simulator of TE sequence evolution, and used it to generate a benchmark with which we evaluated the MSA predictions produced by several popular aligners, along with Refiner, a method we developed in the context of our RepeatModeler software. We find that MAFFT and Refiner generally outperform other aligners for low to medium divergence simulated sequences, while Refiner is uniquely effective when tasked with aligning high-divergent and fragmented instances of a family.

Current progress and open challenges for applying deep learning across the biosciences

Deep Learning (DL) has recently enabled unprecedented advances in one of the grand challenges in computational biology: the half-century-old problem of protein structure prediction. In this paper we discuss recent advances, limitations, and future perspectives of DL on five broad areas: protein structure prediction, protein function prediction, genome engineering, systems biology and data integration, and phylogenetic inference. We discuss each application area and cover the main bottlenecks of DL approaches, such as training data, problem scope, and the ability to leverage existing DL architectures in new contexts. To conclude, we provide a summary of the subject-specific and general challenges for DL across the biosciences.

Deciphering Microbial Gene Family Evolution Using Duplication-Transfer-Loss Reconciliation and RANGER-DTL

Phylogenetic reconciliation has emerged as a principled, highly effective technique for investigating the origin, spread, and evolutionary history of microbial gene families. Proper application of phylogenetic reconciliation requires a clear understanding of potential pitfalls and sources of error, and knowledge of the most effective reconciliation-based tools and protocols to use to maximize accuracy. In this book chapter, we provide a brief overview of Duplication-Transfer-Loss (DTL) reconciliation, the standard reconciliation model used to study microbial gene families and provide a step-by-step computational protocol to maximize the accuracy of DTL reconciliation and minimize false-positive evolutionary inferences.

A probabilistic model for indel evolution: differentiating insertions from deletions

Insertions and deletions (indels) are common molecular evolutionary events. However, probabilistic models for indel evolution are under-developed due to their computational complexity. Here, we introduce several improvements to indel modeling: 1) While previous models for indel evolution assumed that the rates and length distributions of insertions and deletions are equal, here we propose a richer model that explicitly distinguishes between the two; 2) we introduce numerous summary statistics that allow approximate Bayesian computation-based parameter estimation; 3) we develop a method to correct for biases introduced by alignment programs, when inferring indel parameters from empirical data sets; and 4) using a model-selection scheme, we test whether the richer model better fits biological data compared with the simpler model. Our analyses suggest that both our inference scheme and the model-selection procedure achieve high accuracy on simulated data. We further demonstrate that our proposed richer model better fits a large number of empirical data sets and that, for the majority of these data sets, the deletion rate is higher than the insertion rate.

Accurate Large-scale Phylogeny-Aware Alignment using BAli-Phy

MOTIVATION

BAli-Phy, a popular Bayesian method that co-estimates multiple sequence alignments and phylogenetic trees, is a rigorous statistical method, but due to its computational requirements, it has generally been limited to relatively small datasets (at most about 100 sequences). Here we repurpose BAli-Phy as a ``phylogeny-aware" alignment method: we estimate the phylogeny from the input of unaligned sequences, and then use that as a fixed tree within BAli-Phy.

RESULTS

We show that this approach achieves high accuracy, greatly superior to Prank, the current most popular phylogeny-aware alignment method, and is even more accurate than MAFFT, one of the top performing alignment methods in common use. Furthermore, this approach can be used to align very large datasets (up to 1000 sequences in this study).

AVAILABILITY

See https://doi.org/10.13012/B2IDB-7863273_V1 for datasets used in this study.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

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.

Revisiting Evaluation of Multiple Sequence Alignment Methods

Multiple sequence alignment is a core first step in many bioinformatics analyses, and errors in these alignments can have negative consequences for scientific studies. In this article, we review some of the recent literature evaluating multiple sequence alignment methods and identify specific challenges that arise when performing these evaluations. In particular, we discuss the different trends observed in simulation studies and when using biological benchmarks. Overall, we find that multiple sequence alignment, far from being a "solved problem," would benefit from new attention.

Multiple Sequence Alignment for Large Heterogeneous Datasets Using SATé, PASTA, and UPP

The estimation of very large multiple sequence alignments is a challenging problem that requires special techniques in order to achieve high accuracy. Here we describe two software packages-PASTA and UPP-for constructing alignments on large and ultra-large datasets. Both methods have been able to produce highly accurate alignments on 1,000,000 sequences, and trees computed on these alignments are also highly accurate. PASTA provides the best tree accuracy when the input sequences are all full-length, but UPP provides improved accuracy compared to PASTA and other methods when the input contains a large number of fragmentary sequences. Both methods are available in open source form on GitHub.

Intragenic Conflict in Phylogenomic Data Sets

Most phylogenetic analyses assume that a single evolutionary history underlies one gene. However, both biological processes and errors can cause intragenic conflict. The extent to which this conflict is present in empirical data sets is not well documented, but if common, could have far-reaching implications for phylogenetic analyses. We examined several large phylogenomic data sets from diverse taxa using a fast and simple method to identify well-supported intragenic conflict. We found conflict to be highly variable between data sets, from 1% to >92% of genes investigated. We analyzed four exemplar genes in detail and analyzed simulated data under several scenarios. Our results suggest that alignment error may be one major source of conflict, but other conflicts remain unexplained and may represent biological signal or other errors. Whether as part of data analysis pipelines or to explore biologically processes, analyses of within-gene phylogenetic signal should become common.

Improving the thermostability and acid resistance of Rhizopus oryzae α-amylase by using multiple sequence alignment based site-directed mutagenesis

Higher thermostability or acid resistance for fungal α-amylase will help to improve the sugar-making process and cut down the production costs. Here, the thermostability or acid resistance of Rhizopus oryzae α-amylase (ROAmy) was significantly enhanced by site-directed evolution based on multiple sequence alignment (MSA) method. For instance, compared with the wild-type ROAmy, the optimum temperature of mutants G136D and A144Y was increased from 50 to 55 °C, whereas for mutants V174R and I276P, the optimum temperature was increased from 50 to 60 °C. The optimum pH of mutants G136D and A144Y shifted from 5.5 to 5.0, whereas for mutants V174R and T253E, the optimum pH changed from 5.5 to 4.5. The results showed that mutant V174R had a 2.52-fold increase in half-life at 55 °C, a 2.55-fold increase in half-life at pH 4.5, and a 1.61-fold increase in catalytic efficiency (k_cat /K_m ) on soluble starch. The three-dimensional model simulation revealed that changes of hydrophilicity, hydrogen bond, salt bridge, or rigidity observed in mutants might mainly account for the improvement of thermostability and acid resistance. The mutants with improved catalytic properties attained in this work may render an accessible and operable approach for directed evolution of fungal α-amylase aimed at interesting functions.

A new lineage of segmented RNA viruses infecting animals

Metagenomic sequencing has revolutionised our knowledge of virus diversity, with new virus sequences being reported faster than ever before. However, virus discovery from metagenomic sequencing usually depends on detectable homology: without a sufficiently close relative, so-called 'dark' virus sequences remain unrecognisable. An alternative approach is to use virus-identification methods that do not depend on detecting homology, such as virus recognition by host antiviral immunity. For example, virus-derived small RNAs have previously been used to propose 'dark' virus sequences associated with the Drosophilidae (Diptera). Here, we combine published Drosophila data with a comprehensive search of transcriptomic sequences and selected meta-transcriptomic datasets to identify a completely new lineage of segmented positive-sense single-stranded RNA viruses that we provisionally refer to as the Quenyaviruses. Each of the five segments contains a single open reading frame, with most encoding proteins showing no detectable similarity to characterised viruses, and one sharing a small number of residues with the RNA-dependent RNA polymerases of single- and double-stranded RNA viruses. Using these sequences, we identify close relatives in approximately 20 arthropods, including insects, crustaceans, spiders, and a myriapod. Using a more conserved sequence from the putative polymerase, we further identify relatives in meta-transcriptomic datasets from gut, gill, and lung tissues of vertebrates, reflecting infections of vertebrates or of their associated parasites. Our data illustrate the utility of small RNAs to detect viruses with limited sequence conservation, and provide robust evidence for a new deeply divergent and phylogenetically distinct RNA virus lineage.

Exploring the impact of morphology, multiple sequence alignment and choice of optimality criteria in phylogenetic inference: a case study with the Neotropical orb-weaving spider genus Wagneriana (Araneae: Araneidae)

We present a total evidence phylogenetic analysis of the Neotropical orb-weaving spider genus Wagneriana and discuss the phylogenetic impacts of methodological choices. We analysed 167 phenotypic characters and nine loci scored for 115 Wagneriana and outgroups, including 46 newly sequenced species. We compared total evidence analyses and molecular-only analyses to evaluate the impact of phenotypic evidence, and we performed analyses using the programs POY, TNT, RAxML, GARLI, IQ-TREE and MrBayes to evaluate the effects of multiple sequence alignment and optimality criteria. In all analyses, Wagneriana carimagua and Wagneriana uropygialis were nested in the genera Parawixia and Alpaida, respectively, and the remaining species of Wagneriana fell into three main clades, none of which formed a pair of sister taxa. However, sister-group relationships among the main clades and their internal relationships were strongly influenced by methodological choices. Alignment methods had comparable topological effects to those of optimality criteria in terms of ‘subtree pruning and regrafting’ moves. The inclusion of phenotypic evidence, 2.80–3.05% of the total evidence matrices, increased support irrespective of the optimality criterion used. The monophyly of some groups was recovered only after the addition of morphological characters. A new araneid genus, Popperaneus gen. nov., is erected, and Paraverrucosa is resurrected. Four new synonymies and seven new combinations are proposed.

P-element Somatic Inhibitor Protein Binding a Target Sequence in dsx Pre-mRNA Conserved in Bombyx mori and Spodoptera litura

Bombyx mori doublesex (Bmdsx) functions as a double-switch gene in the final step of the sex-determination cascade in the silkworm Bombyx mori. The P-element somatic inhibitor (PSI) protein in B. mori interacts with Bmdsx pre-mRNA in CE1 as an exonic splicing silencer to promote male-specific splicing of Bmdsx. However, the character of the interaction between BmPSI and Bmdsx pre-mRNA remains unclear. Electrophoretic mobility shift assay (EMSA) results showed that the four KH_1 motifs in BmPSI are all essential for the binding, especially the former two KH_1 motifs. Three active sites (I116, L127, and IGGI) in the KH_1 motif were found to be necessary for the binding through EMSA, circular dichroism (CD) spectroscopy, and isothermal titration calorimetry (ITC). The PSI homologous protein in S. litura (SlPSI) was purified and the binding of SlPSI and CE1 was verified. Compared with BmPSI, the mutant SlPSI proteins of I116 and IGGI lost their ability to bind to CE1. In conclusion, the binding of PSI and dsx pre-mRNA are generally conserved in both B. mori and S. litura. These findings provide clues for sex determination in Lepidoptera.