Homology Inference Research Articles

Incorporating Nearest-Neighbor Site Dependence into Protein Evolution Models.

Evolutionary models of proteins are widely used for statistical sequence alignment and inference of homology and phylogeny. However, the vast majority of these models rely on an unrealistic assumption of independent evolution between sites. Here we focus on the related problem of protein structure alignment, a classic tool of computational biology that is widely used to identify structural and functional similarity and to infer homology among proteins. A site-independent statistical model for protein structural evolution has previously been introduced and shown to significantly improve alignments and phylogenetic inferences compared with approaches that utilize only amino acid sequence information. Here we extend this model to account for correlated evolutionary drift among neighboring amino acid positions. The result is a spatiotemporal model of protein structure evolution, described by a multivariate diffusion process convolved with a spatial birth-death process. This extended site-dependent model (SDM) comes with little additional computational cost or analytical complexity compared with the site-independent model (SIM). We demonstrate that this SDM yields a significant reduction of bias in estimated evolutionary distances and helps further improve phylogenetic tree reconstruction. We also develop a simple model of site-dependent sequence evolution, which we use to demonstrate the bias resulting from the application of standard site-independent sequence evolution models.

Sequential Paleotetraploidization shaped the carrot genome

BackgroundCarrot (Daucus carota subsp. carota L.) is an important root crop with an available high-quality genome. The carrot genome is thought to have undergone recursive paleo-polyploidization, but the extent, occurrences, and nature of these events are not clearly defined.ResultsUsing a previously published comparative genomics pipeline, we reanalysed the carrot genome and characterized genomic fractionation, as well as gene loss and retention, after each of the two tetraploidization events and inferred a dominant and sensitive subgenome for each event. In particular, we found strong evidence of two sequential tetraploidization events, with one (Dc-α) approximately 46–52 million years ago (Mya) and the other (Dc-β) approximately 77–87 Mya, both likely allotetraploidization in nature. The Dc-β event was likely common to all Apiales plants, occurring around the divergence of Apiales-Bruniales and after the divergence of Apiales-Asterales, likely playing an important role in the derivation and divergence of Apiales species. Furthermore, we found that rounds of polyploidy events contributed to the expansion of gene families responsible for plastidial methylerythritol phosphate (MEP), the precursor of carotenoid accumulation, and shaped underlying regulatory pathways. The alignment of orthologous and paralogous genes related to different events of polyploidization and speciation constitutes a comparative genomics platform for studying Apiales, Asterales, and many other related species.ConclusionsHierarchical inference of homology revealed two tetraploidization events that shaped the carrot genome, which likely contributed to the successful establishment of Apiales plants and the expansion of MEP, upstream of the carotenoid accumulation pathway.

A mass spectrometry guided approach for the identification of novel vaccine candidates in gram-negative pathogens

Vaccination is the most effective method to prevent infectious diseases. However, approaches to identify novel vaccine candidates are commonly laborious and protracted. While surface proteins are suitable vaccine candidates and can elicit antibacterial antibody responses, systematic approaches to define surfomes from gram-negatives have rarely been successful. Here we developed a combined discovery-driven mass spectrometry and computational strategy to identify bacterial vaccine candidates and validate their immunogenicity using a highly prevalent gram-negative pathogen, Helicobacter pylori, as a model organism. We efficiently isolated surface antigens by enzymatic cleavage, with a design of experiment based strategy to experimentally dissect cell surface-exposed from cytosolic proteins. From a total of 1,153 quantified bacterial proteins, we thereby identified 72 surface exposed antigens and further prioritized candidates by computational homology inference within and across species. We next tested candidate-specific immune responses. All candidates were recognized in sera from infected patients, and readily induced antibody responses after vaccination of mice. The candidate jhp_0775 induced specific B and T cell responses and significantly reduced colonization levels in mouse therapeutic vaccination studies. In infected humans, we further show that jhp_0775 is immunogenic and activates IFNγ secretion from peripheral CD4+ and CD8+ T cells. Our strategy provides a generic preclinical screening, selection and validation process for novel vaccine candidates against gram-negative bacteria, which could be employed to other gram-negative pathogens.

Phylostratr: a framework for phylostratigraphy.

The goal of phylostratigraphy is to infer the evolutionary origin of each gene in an organism. This is done by searching for homologs within increasingly broad clades. The deepest clade that contains a homolog of the protein(s) encoded by a gene is that gene's phylostratum. We have created a general R-based framework, phylostratr, to estimate the phylostratum of every gene in a species. The program fully automates analysis: selecting species for balanced representation, retrieving sequences, building databases, inferring phylostrata and returning diagnostics. Key diagnostics include: detection of genes with inferred homologs in old clades, but not intermediate ones; proteome quality assessments; false-positive diagnostics, and checks for missing organellar genomes. phylostratr allows extensive customization and systematic comparisons of the influence of analysis parameters or genomes on phylostrata inference. A user may: modify the automatically generated clade tree or use their own tree; provide custom sequences in place of those automatically retrieved from UniProt; replace BLAST with an alternative algorithm; or tailor the method and sensitivity of the homology inference classifier. We show the utility of phylostratr through case studies in Arabidopsis thaliana and Saccharomyces cerevisiae. Source code available at https://github.com/arendsee/phylostratr. Supplementary data are available at Bioinformatics online.

Variable sources of Bk virus in renal allograft recipients

BK virus is the causative agent of polyomavirus-associated nephropathy, a major cause of kidney transplant failure affecting 1%-10% of recipients. Previous studies that investigated the viral source on the kidney recipient pointed that the donor is implicated in the origin of human polyomavirus BK (BKPyV) infection in recipients, but giving the low genetic variability of BKPyV this subject is still controversial. The aim of this study was to determine if BKPyV replicating in kidney recipients after transplantation is always originated from the donor. Urine and blood samples from 68 pairs of living donors and kidney recipients who underwent renal transplantation from August 2010-September 2011 were screened for BKPyV by real time polymerase chain reaction. Only three recipients presented viremia. When both donors and recipients were BKPyV positive, a larger fragment of VP1 region was obtained and sequenced to determine the level of similarity between them. A phylogenetic tree was built for the 12 pairs of sequences obtained from urine and high level of similarity among all sequences was observed, indicating that homology inferences for donor and recipient viruses must be cautiously interpreted. However, a close inspection on the donor-recipient pairs sequences revealed that 3 of 12 pairs presented considerably different viruses and 4 of 12 presented mixed infection, indicating that the source of BKPyV infection is not exclusively derived from the donor. We report that about 60% of the renal recipients shed BKPyV genetically distinct from the donor, confronting the accepted concept that the donor is the main source of recipients' infection.

Efficient inference of homologs in large eukaryotic pan-proteomes

BackgroundIdentification of homologous genes is fundamental to comparative genomics, functional genomics and phylogenomics. Extensive public homology databases are of great value for investigating homology but need to be continually updated to incorporate new sequences. As new sequences are rapidly being generated, there is a need for efficient standalone tools to detect homologs in novel data.ResultsTo address this, we present a fast method for detecting homology groups across a large number of individuals and/or species. We adopted a k-mer based approach which considerably reduces the number of pairwise protein alignments without sacrificing sensitivity. We demonstrate accuracy, scalability, efficiency and applicability of the presented method for detecting homology in large proteomes of bacteria, fungi, plants and Metazoa.ConclusionsWe clearly observed the trade-off between recall and precision in our homology inference. Favoring recall or precision strongly depends on the application. The clustering behavior of our program can be optimized for particular applications by altering a few key parameters. The program is available for public use at https://github.com/sheikhizadeh/pantools as an extension to our pan-genomic analysis tool, PanTools.

Organ Homologies and Perianth Evolution in the Dasymaschalon Alliance (Annonaceae): Inner Petal Loss and Its Functional Consequences.

The Dasymaschalon alliance within the early divergent angiosperm family Annonaceae comprises c. 180 species in four genera (Dasymaschalon, Desmos, Friesodielsia, and Monanthotaxis). The alliance offers an excellent opportunity for investigating perianth evolution and functional adaptations because of the presence of different numbers of petal whorls and contrasting floral chamber morphologies. The absence of the inner petal whorl in Dasymaschalon renders it distinctive in the family: previous studies have suggested that its three outermost stamens might be homologous with the inner petals of the sister genus, Friesodielsia, reflecting a homeotic shift of floral organ identify from inner petals to stamens. To investigate this hypothesis and general perianth evolution in the alliance, we (i) compared the floral vascularization of selected Dasymaschalon and Friesodielsia species using paraffin serial sectioning, and (ii) mapped selected perianth characters of inferred functional significance onto a molecular phylogenetic framework of the Dasymaschalon alliance (46 accessions; five cpDNA, and two nrDNA markers). The results indicate that the vasculature of the outermost stamen whorl of Dasymaschalon does not fuse with the perianth cortical vascular system, but instead splits from the basal traces of the free stamen bundles, contradicting previous inferences of homology with the inner corolla whorl of other Annonaceae. The loss of the inner petal whorl in Dasymaschalon is less likely to be due to a homeotic mutation, and instead possibly involved either the loss of genes that are responsible for determining inner petals or else the expression failure of these genes. Optimizations of perianth characters indicate that the absence of the inner petal whorl and the connivence of outer petals during anthesis are synapomorphic for Dasymaschalon. Circadian trapping of pollinators is inferred either to be derived in the stem lineage of the Dasymaschalon–Friesodielsia clade, or else to have evolved in parallel in the Dasymaschalon and Friesodielsia lineages. Subsequent changes in the remaining petals of Dasymaschalon flowers (which do not fully separate during anthesis) are likely to have enabled perpetuation of the circadian trapping mechanism, lessening the adverse impacts of inner petal loss.

Homology Inference Based on a Reconciliation Approach for the Comparative Genomics of Protozoa.

Protozoa parasites are responsible for several diseases in tropical countries, such as malaria, sleeping sickness, Chagas disease, leishmaniasis, amebiasis, and giardiasis, which together threaten millions of people around the world. In addition, most of the classic parasitic diseases due to protozoa are zoonotic. Understanding the biology of these organisms plays a relevant role in combating these diseases. Using homology inference and comparative genomics, this study targeted 3 protozoan species from different Phyla: Cryptosporidium muris (Apicomplexa), Entamoeba invadens (Amoebozoa), and Trypanosoma grayi (Euglenozoa). In this study, we propose a new approach for the identification of homologs, based on the reconciliation of the results of 2 different homology inference software programs. Our results showed that 46.1% (59/128) of the groups inferred by our reconciliation approach could be validated using this methodology. These validated groups are here called homologous Supergroups and were compared with SUPERFAMILY and Pfam Clans.

An Overlooked Paleotetraploidization in Cucurbitaceae.

Cucurbitaceae plants are of considerable biological and economic importance, and genomes of cucumber, watermelon, and melon have been sequenced. However, a comparative genomics exploration of their genome structures and evolution has not been available. Here, we aimed at performing a hierarchical inference of genomic homology resulted from recursive paleopolyploidizations. Unexpectedly, we found that, shortly after a core-eudicot-common hexaploidy, a cucurbit-common tetraploidization (CCT) occurred, overlooked by previous reports. Moreover, we characterized gene loss (and retention) after these respective events, which were significantly unbalanced between inferred subgenomes, and between plants after their split. The inference of a dominant subgenome and a sensitive one suggested an allotetraploid nature of the CCT. Besides, we found divergent evolutionary rates among cucurbits, and after doing rate correction, we dated the CCT to be 90–102 Ma, likely common to all Cucurbitaceae plants, showing its important role in the establishment of the plant family.

Improved transcriptome sampling pinpoints 26 ancient and more recent polyploidy events in Caryophyllales, including two allopolyploidy events.

Studies of the macroevolutionary legacy of polyploidy are limited by an incomplete sampling of these events across the tree of life. To better locate and understand these events, we need comprehensive taxonomic sampling as well as homology inference methods that accurately reconstruct the frequency and location of gene duplications. We assembled a data set of transcriptomes and genomes from 168 species in Caryophyllales, of which 43 transcriptomes were newly generated for this study, representing one of the most densely sampled genomic-scale data sets available. We carried out phylogenomic analyses using a modified phylome strategy to reconstruct the species tree. We mapped the phylogenetic distribution of polyploidy events by both tree-based and distance-based methods, and explicitly tested scenarios for allopolyploidy. We identified 26 ancient and more recent polyploidy events distributed throughout Caryophyllales. Two of these events were inferred to be allopolyploidy. Through dense phylogenomic sampling, we show the propensity of polyploidy throughout the evolutionary history of Caryophyllales. We also provide a framework for utilizing transcriptome data to detect allopolyploidy, which is important as it may have different macroevolutionary implications compared with autopolyploidy.

Phylogenetic relationships among tribes of the green lacewing subfamily Chrysopinae recovered based on mitochondrial phylogenomics

Chrysopidae (green lacewings) is the second largest family in Neuroptera, and it includes medium-size lacewings largely recognized by the presence of golden-colored eyes, bright green bodies and delicate wings with dense venation patterns. The subfamily Chrysopinae includes 97% of the species diversity in the family and it is currently divided into four tribes: Ankylopterygini, Belonopterygini, Chrysopini and Leucochrysini. Here we sequenced and annotated the nearly complete mitochondrial genomes of four species of each these tribes: Abachrysa eureka, Italochrysa insignis, Leucochrysa pretiosa, Parankyloteryx sp. We then reconstructed the phylogenetic relationships with estimated divergence times among tribes of Chrysopinae based on the mt genomic data. Our results suggest that Chrysopinae sans Nothancyla verreauxi evolved as two reciprocally monophyletic lineages formed by stem members of the tribes Leucochrysini plus Belonopterygini on one hand, and the stem members of Ankylopterygini plus Chrysopini on the other. Our estimations of divergence times place the diversification of stem Chrysopinae into the extant tribes during the Middle Jurassic to Late Cretaceous. The relatively young ages previously estimated for the green lacewing divergences were probably underestimated due to false inferences of homology between non-sister taxa that are later correctly identified as homoplasy after more taxa are added.

ProtozoaDB 2.0: A Trypanosoma Brucei Case Study.

Over the last decade new species of Protozoa have been sequenced and deposited in GenBank. Analyzing large amounts of genomic data, especially using Next Generation Sequencing (NGS), is not a trivial task, considering that researchers used to deal or focus their studies on few genes or gene families or even small genomes. To facilitate the information extraction process from genomic data, we developed a database system called ProtozoaDB that included five genomes of Protozoa in its first version. In the present study, we present a new version of ProtozoaDB called ProtozoaDB 2.0, now with the genomes of 22 pathogenic Protozoa. The system has been fully remodeled to allow for new tools and a more expanded view of data, and now includes a number of analyses such as: (i) similarities with other databases (model organisms, the Conserved Domains Database, and the Protein Data Bank); (ii) visualization of KEGG metabolic pathways; (iii) the protein structure from PDB; (iv) homology inferences; (v) the search for related publications in PubMed; (vi) superfamily classification; and (vii) phenotype inferences based on comparisons with model organisms. ProtozoaDB 2.0 supports RESTful Web Services to make data access easier. Those services were written in Ruby language using Ruby on Rails (RoR). This new version also allows a more detailed analysis of the object of study, as well as expanding the number of genomes and proteomes available to the scientific community. In our case study, a group of prenyltransferase proteinsalready described in the literature was found to be a good drug target for Trypanosomatids.

Dynamical persistence of active sites identified in maltose-binding protein.

This study identifies dynamical properties of maltose-binding protein (MBP) useful in unveiling active site residues susceptible to ligand binding. The described methodology has been previously used in support of novel topological techniques of persistent homology and statistical inference in complex, multi-scale, high-dimensional data often encountered in computational biophysics. Here we outline a computational protocol that is based on the anisotropic elastic network models of 14 all-atom three-dimensional protein structures. We introduce the notion of dynamical distance matrices as a measure of correlated interactions among 370 amino acid residues that constitute a single protein. The dynamical distance matrices serve as an input for a persistent homology suite of codes to further distinguish a small subset of residues with high affinity for ligand binding and allosteric activity. In addition, we show that ligand-free closed MBP structures require lower deformation energies than open MBP structures, which may be used in categorization of time-evolving molecular dynamics structures. Analysis of the most probable allosteric coupling pathways between active site residues and the protein exterior is also presented.

GenFamClust: an accurate, synteny-aware and reliable homology inference algorithm.

BackgroundHomology inference is pivotal to evolutionary biology and is primarily based on significant sequence similarity, which, in general, is a good indicator of homology. Algorithms have also been designed to utilize conservation in gene order as an indication of homologous regions. We have developed GenFamClust, a method based on quantification of both gene order conservation and sequence similarity.ResultsIn this study, we validate GenFamClust by comparing it to well known homology inference algorithms on a synthetic dataset. We applied several popular clustering algorithms on homologs inferred by GenFamClust and other algorithms on a metazoan dataset and studied the outcomes. Accuracy, similarity, dependence, and other characteristics were investigated for gene families yielded by the clustering algorithms. GenFamClust was also applied to genes from a set of complete fungal genomes and gene families were inferred using clustering. The resulting gene families were compared with a manually curated gold standard of pillars from the Yeast Gene Order Browser. We found that the gene-order component of GenFamClust is simple, yet biologically realistic, and captures local synteny information for homologs.ConclusionsThe study shows that GenFamClust is a more accurate, informed, and comprehensive pipeline to infer homologs and gene families than other commonly used homology and gene-family inference methods.Electronic supplementary materialThe online version of this article (doi:10.1186/s12862-016-0684-2) contains supplementary material, which is available to authorized users.

Benchmarking the next generation of homology inference tools.

Motivation: Over the last decades, vast numbers of sequences were deposited in public databases. Bioinformatics tools allow homology and consequently functional inference for these sequences. New profile-based homology search tools have been introduced, allowing reliable detection of remote homologs, but have not been systematically benchmarked. To provide such a comparison, which can guide bioinformatics workflows, we extend and apply our previously developed benchmark approach to evaluate the ‘next generation’ of profile-based approaches, including CS-BLAST, HHSEARCH and PHMMER, in comparison with the non-profile based search tools NCBI-BLAST, USEARCH, UBLAST and FASTA.Method: We generated challenging benchmark datasets based on protein domain architectures within either the PFAM + Clan, SCOP/Superfamily or CATH/Gene3D domain definition schemes. From each dataset, homologous and non-homologous protein pairs were aligned using each tool, and standard performance metrics calculated. We further measured congruence of domain architecture assignments in the three domain databases.Results: CSBLAST and PHMMER had overall highest accuracy. FASTA, UBLAST and USEARCH showed large trade-offs of accuracy for speed optimization.Conclusion: Profile methods are superior at inferring remote homologs but the difference in accuracy between methods is relatively small. PHMMER and CSBLAST stand out with the highest accuracy, yet still at a reasonable computational cost. Additionally, we show that less than 0.1% of Swiss-Prot protein pairs considered homologous by one database are considered non-homologous by another, implying that these classifications represent equivalent underlying biological phenomena, differing mostly in coverage and granularity.Availability and Implementation: Benchmark datasets and all scripts are placed at (http://sonnhammer.org/download/Homology_benchmark).Contact: forslund@embl.deSupplementary information: Supplementary data are available at Bioinformatics online.

Parameterizing sequence alignment with an explicit evolutionary model.

BackgroundInference of sequence homology is inherently an evolutionary question, dependent upon evolutionary divergence. However, the insertion and deletion penalties in the most widely used methods for inferring homology by sequence alignment, including BLAST and profile hidden Markov models (profile HMMs), are not based on any explicitly time-dependent evolutionary model. Using one fixed score system (BLOSUM62 with some gap open/extend costs, for example) corresponds to making an unrealistic assumption that all sequence relationships have diverged by the same time. Adoption of explicit time-dependent evolutionary models for scoring insertions and deletions in sequence alignments has been hindered by algorithmic complexity and technical difficulty.ResultsWe identify and implement several probabilistic evolutionary models compatible with the affine-cost insertion/deletion model used in standard pairwise sequence alignment. Assuming an affine gap cost imposes important restrictions on the realism of the evolutionary models compatible with it, as single insertion events with geometrically distributed lengths do not result in geometrically distributed insert lengths at finite times. Nevertheless, we identify one evolutionary model compatible with symmetric pair HMMs that are the basis for Smith-Waterman pairwise alignment, and two evolutionary models compatible with standard profile-based alignment.We test different aspects of the performance of these “optimized branch length” models, including alignment accuracy and homology coverage (discrimination of residues in a homologous region from nonhomologous flanking residues). We test on benchmarks of both global homologies (full length sequence homologs) and local homologies (homologous subsequences embedded in nonhomologous sequence).ConclusionsContrary to our expectations, we find that for global homologies a single long branch parameterization suffices both for distant and close homologous relationships. In contrast, we do see an advantage in using explicit evolutionary models for local homologies. Optimal branch parameterization reduces a known artifact called “homologous overextension”, in which local alignments erroneously extend through flanking nonhomologous residues.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0832-5) contains supplementary material, which is available to authorized users.

Improved orthologous databases to ease protozoan targets inference.

BackgroundHomology inference helps on identifying similarities, as well as differences among organisms, which provides a better insight on how closely related one might be to another. In addition, comparative genomics pipelines are widely adopted tools designed using different bioinformatics applications and algorithms. In this article, we propose a methodology to build improved orthologous databases with the potential to aid on protozoan target identification, one of the many tasks which benefit from comparative genomics tools.MethodsOur analyses are based on OrthoSearch, a comparative genomics pipeline originally designed to infer orthologs through protein-profile comparison, supported by an HMM, reciprocal best hits based approach. Our methodology allows OrthoSearch to confront two orthologous databases and to generate an improved new one. Such can be later used to infer potential protozoan targets through a similarity analysis against the human genome.ResultsThe protein sequences of Cryptosporidium hominis, Entamoeba histolytica and Leishmania infantum genomes were comparatively analyzed against three orthologous databases: (i) EggNOG KOG, (ii) ProtozoaDB and (iii) Kegg Orthology (KO). That allowed us to create two new orthologous databases, “KO + EggNOG KOG” and “KO + EggNOG KOG + ProtozoaDB”, with 16,938 and 27,701 orthologous groups, respectively.Such new orthologous databases were used for a regular OrthoSearch run. By confronting “KO + EggNOG KOG” and “KO + EggNOG KOG + ProtozoaDB” databases and protozoan species we were able to detect the following total of orthologous groups and coverage (relation between the inferred orthologous groups and the species total number of proteins): Cryptosporidium hominis: 1,821 (11 %) and 3,254 (12 %); Entamoeba histolytica: 2,245 (13 %) and 5,305 (19 %); Leishmania infantum: 2,702 (16 %) and 4,760 (17 %).Using our HMM-based methodology and the largest created orthologous database, it was possible to infer 13 orthologous groups which represent potential protozoan targets; these were found because of our distant homology approach.We also provide the number of species-specific, pair-to-pair and core groups from such analyses, depicted in Venn diagrams.ConclusionsThe orthologous databases generated by our HMM-based methodology provide a broader dataset, with larger amounts of orthologous groups when compared to the original databases used as input. Those may be used for several homology inference analyses, annotation tasks and protozoan targets identification.Electronic supplementary materialThe online version of this article (doi:10.1186/s13071-015-1090-0) contains supplementary material, which is available to authorized users.

DissectHMMER: a HMMER-based score dissection framework that statistically evaluates fold-critical sequence segments for domain fold similarity.

BackgroundAnnotation transfer for function and structure within the sequence homology concept essentially requires protein sequence similarity for the secondary structural blocks forming the fold of a protein. A simplistic similarity approach in the case of non-globular segments (coiled coils, low complexity regions, transmembrane regions, long loops, etc.) is not justified and a pertinent source for mistaken homologies. The latter is either due to positional sequence conservation as a result of a very simple, physically induced pattern or integral sequence properties that are critical for function. Furthermore, against the backdrop that the number of well-studied proteins continues to grow at a slow rate, it necessitates for a search methodology to dive deeper into the sequence similarity space to connect the unknown sequences to the well-studied ones, albeit more distant, for biological function postulations.ResultsBased on our previous work of dissecting the hidden markov model (HMMER) based similarity score into fold-critical and the non-globular contributions to improve homology inference, we propose a framework-dissectHMMER, that identifies more fold-related domain hits from standard HMMER searches. Subsequent statistical stratification of the fold-related hits into cohorts of functionally-related domains allows for the function postulation of the query sequence. Briefly, the technical problems as to how to recognize non-globular parts in the domain model, resolve contradictory HMMER2/HMMER3 results and evaluate fold-related domain hits for homology, are addressed in this work. The framework is benchmarked against a set of SCOP-to-Pfam domain models. Despite being a sequence-to-profile method, dissectHMMER performs favorably against a profile-to-profile based method-HHsuite/HHsearch. Examples of function annotation using dissectHMMER, including the function discovery of an uncharacterized membrane protein Q9K8K1_BACHD (WP_010899149.1) as a lactose/H+ symporter, are presented. Finally, dissectHMMER webserver is made publicly available at http://dissecthmmer.bii.a-star.edu.sg.ConclusionsThe proposed framework-dissectHMMER, is faithful to the original inception of the sequence homology concept while improving upon the existing HMMER search tool through the rescue of statistically evaluated false-negative yet fold-related domain hits to the query sequence. Overall, this translates into an opportunity for any novel protein sequence to be functionally characterized.ReviewersThis article was reviewed by Masanori Arita, Shamil Sunyaev and L. Aravind.Electronic supplementary materialThe online version of this article (doi:10.1186/s13062-015-0068-3) contains supplementary material, which is available to authorized users.

Evaluation of BLAST-based edge-weighting metrics used for homology inference with the Markov Clustering algorithm.

BackgroundClustering protein sequences according to inferred homology is a fundamental step in the analysis of many large data sets. Since the publication of the Markov Clustering (MCL) algorithm in 2002, it has been the centerpiece of several popular applications. Each of these approaches generates an undirected graph that represents sequences as nodes connected to each other by edges weighted with a BLAST-based metric. MCL is then used to infer clusters of homologous proteins by analyzing these graphs. The various approaches differ only by how they weight the edges, yet there has been very little direct examination of the relative performance of alternative edge-weighting metrics. This study compares the performance of four BLAST-based edge-weighting metrics: the bit score, bit score ratio (BSR), bit score over anchored length (BAL), and negative common log of the expectation value (NLE). Performance is tested using the Extended CEGMA KOGs (ECK) database, which we introduce here.ResultsAll metrics performed similarly when analyzing full-length sequences, but dramatic differences emerged as progressively larger fractions of the test sequences were split into fragments. The BSR and BAL successfully rescued subsets of clusters by strengthening certain types of alignments between fragmented sequences, but also shifted the largest correct scores down near the range of scores generated from spurious alignments. This penalty outweighed the benefits in most test cases, and was greatly exacerbated by increasing the MCL inflation parameter, making these metrics less robust than the bit score or the more popular NLE. Notably, the bit score performed as well or better than the other three metrics in all scenarios.ConclusionsThe results provide a strong case for use of the bit score, which appears to offer equivalent or superior performance to the more popular NLE. The insight that MCL-based clustering methods can be improved using a more tractable edge-weighting metric will greatly simplify future implementations. We demonstrate this with our own minimalist Python implementation: Porthos, which uses only standard libraries and can process a graph with 25 m + edges connecting the 60 k + KOG sequences in half a minute using less than half a gigabyte of memory.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0625-x) contains supplementary material, which is available to authorized users.

Using homology relations within a database markedly boosts protein sequence similarity search

Inference of homology from protein sequences provides an essential tool for analyzing protein structure, function, and evolution. Current sequence-based homology search methods are still unable to detect many similarities evident from protein spatial structures. In computer science a search engine can be improved by considering networks of known relationships within the search database. Here, we apply this idea to protein-sequence-based homology search and show that it dramatically enhances the search accuracy. Our new method, COMPADRE (COmparison of Multiple Protein sequence Alignments using Database RElationships) assesses the relationship between the query sequence and a hit in the database by considering the similarity between the query and hit's known homologs. This approach increases detection quality, boosting the precision rate from 18% to 83% at half-coverage of all database homologs. The increased precision rate allows detection of a large fraction of protein structural relationships, thus providing structure and function predictions for previously uncharacterized proteins. Our results suggest that this general approach is applicable to a wide variety of methods for detection of biological similarities. The web server is available at prodata.swmed.edu/compadre.