HMMER web server: 2015 update.
Finn RD, Clements J, Arndt W, Miller BL, Wheeler TJ, Schreiber F, Bateman A, Eddy SR. · 2015
Read HMMER web server: 2015 update. by Finn RD, Clements J, Arndt W, Miller BL, Wheeler TJ, Schreiber F, Bateman A, Eddy SR. (2015) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4489315.txt.
Introduction
Robert Finn and seven co-authors updated the HMMER web server in Nucleic Acids Research in 2015. HMMER searches protein sequence databases with profile models, statistical descriptions of protein families, to find relatives of a query sequence and to annotate it with known domains. The website, launched in 2011, brings those algorithms to users without local computing power.
The update adds the iterative search program jackhmmer, which refines a search in rounds. Because databases keep growing, result tables can overwhelm. Hits can now be summarized by the taxonomy of the organisms they come from, or by their domain layout. They can also be filtered before searching by taxonomic group, which speeds the search. The library of profile models grew to include several more curated collections, from CATH-Gene3D to TIGRFAMs. The article also notes the site's relocation to a new address at the European Bioinformatics Institute.
Catalog record
| CREATOR | Finn RD, Clements J, Arndt W, Miller BL, Wheeler TJ, Schreiber F, Bateman A, Eddy SR. |
|---|---|
| YEAR | 2015 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
HMMER web server: 2015 update By Robert D. Finn, Jody Clements, William Arndt, Benjamin L. Miller, Travis J. Wheeler, Fabian Schreiber, Alex Bateman, Sean R. Eddy Published in: Nucleic Acids Research DOI: 10.1093/nar/gkv397 ABSTRACT The HMMER website, available at http://www.ebi.ac.uk/Tools/hmmer/, provides access to the protein homology search algorithms found in the HMMER software suite. Since the first release of the website in 2011, the search repertoire has been expanded to include the iterative search algorithm, jackhmmer. The continued growth of the target sequence databases means that traditional tabular representations of significant sequence hits can be overwhelming to the user. Consequently, additional ways of presenting homology search results have been developed, allowing them to be summarised according to taxonomic distribution or domain architecture. The taxonomy and domain architecture representations can be used in combination to filter the results according to the needs of a user. Searches can also be restricted prior to submission using a new taxonomic filter, which not only ensures that the results are specific to the requested taxonomic group, but also improves search performance. The repertoire of profile hidden Markov model libraries, which are used for annotation of query sequences with protein families and domains, has been expanded to include the libraries from CATH-Gene3D, PIRSF, Superfamily and TIGRFAMs. Finally, we discuss the relocation of the HMMER webserver to the European Bioinformatics Institute and the potential impact that this will have. INTRODUCTION Homology searches are widely used within molecular biology, facilitating the transfer of annotation from a functionally characterised sequence or region to a corresponding region in another sequence. When searching against sequence databases, the HMMER software uses profile hidden Markov models (HMMs) to represent the query (1,2)—which can take the form of a single protein sequence or a multiple sequence alignment. In the case of a multiple sequence alignment, the observed amino acid frequencies in each column are converted to position-specific probabilities, with per position probabilities for both insertions and deletions, determined from the input alignment (1,2). For single sequence searches, a profile HMM is constructed from the sequence using position-independent affine gap open and extension probabilities (defaults: 0.02 and 0.4) and emission probabilities obtained from the inferred probabilistic basis of a standard substitution matrix (default: BLOSUM62) (3,4). In early 2011, the functionality of the website hosting the HMMER software (http://hmmer.org) was expanded to allow online searches of protein sequences against either a protein sequence database or a HMM library (5). This search service not only took advantage of the speed improvements of HMMER3 software (6), but also hardware, the latest approaches to website design and other technical implementations (e.g. in-memory databases and use of NoSQL). The combination of these four aspects allowed the searching of sequences against large sequence databases such as UniProtKB, at near interactive speeds. Websites with minimal loading times (<10 s) are recognised for not interrupting the user's train of thought, and hence increase user productivity (7–9). Since the initial release, the popularity of online HMMER searches has grown, with millions of sequence searches performed per year (averaging over 5200 searches per day or one search every 6 s, search statistics for http://hmmer.org only). These searches are split between requests coming from a browser (20% of searches) or via programmatic access using the RESTful application program interface (API, 80% of searches). For example, RSCB-PDB (10) uses the API to annotate newly deposited structures with Pfam annotations. The initial release of the website provided the following three search algorithms: phmmer—single protein sequence against protein sequence databasehmmscan—single protein sequence against profile HMM library (Pfam)hmmsearch—either multiple sequence alignment or profile HMM against protein sequence database In this article, we describe the recent developments of the website, which include iterative searches with jackhmmer - an expanded repertoire of HMM libraries and a variety of result visualisations that allow rapid interrogation and interpretation of the results. Such developments are set against a constant background of target database growth, which for the large sequence databases results in a constant increase in the dynamic range of hits that will be returned from a query. We demonstrate different ways that searches can be limited and how the results can be progressively dissected according to taxonomy and/or domain architecture (the order of the domain(s) on a sequence). HMMER WEBSERVER DEVELOPMENTS The HMMER webserver has been upgraded to the latest version of the HMMER software, version 3.1b2. This software version includes minor bug fixes, but more importantly has additional performance improvements over the previous version, version 3.0. While this has made the searches faster, the underlying sequence databases have grown substantially. UniProtKB (11) currently contains 91 408 504 sequences (release 2015_02), compared with 13 593 921 in January 2011 (release 2011_01), an increase of 570%. One of the major speed optimisations utilised by the site requires the caching of the sequence database in memory. Due to exponential growth of the databases, it is no longer possible to support both the NCBI non-redundant protein sequence database and UniProtKB. Future software developments will ultimately allow the sharding of the databases to allow future scalability, but this has yet to be implemented within the HMMER daemon (hmmpgmd). Consequently, the website now focuses primarily on UniProtKB (11) as it represents the world's pre-eminent protein database, with sequences annotated either by expert curation or by the application of expert curated rules for automated annotation. While searches remain fast, returning in a matter of a few seconds, subsets of UniProtKB have been included to provide either the highest quality (UniProtKB/Swiss-Prot) annotations, representative sets to provide good coverage of sequence space while reducing the number of potential matches (UniProt Reference Proteomes (11) and Representative Proteome sets (12)) or for curation purposes (pfamseq—Pfam's underlying sequence database). While these subsets do not change the amount of data cached in memory, the smaller target databases increase search performance and make results more manageable. We also include sequences from known structures that have been deposited in the Protein Data Bank (13)). The growth of UniProtKB has been substantial, but it is important to remember that increasing fractions of the new sequences are either identical or nearly identical (>95% identity) to a sequence that already exists in the database (11). Given the nature of this sequence database growth, it is unsurprising that the number of sequences that a query may match from a homology search has equally grown. Thus, many of the developments of the website have focused on trying to improve results visualisation, from summaries to alternative representations to filtering. Expanded results visualisations User experience testing and web usage statistics indicate that it is very difficult to predict what a user is trying to achieve from a homology search. A user's purpose for the search can range from functional annotation to establishing taxonomy distribution to understanding residue conservation in a collection of aligned sequences. The developments described in the following sections have all been designed to enhance access and understanding of the results, whilst catering to a wide range of use cases or enquiries. Sequence matches and features When performing a phmmer search, the single sequence query is also automatically searched against the Pfam (14) profile HMM library using hmmscan, to identify the presence of any Pfam families on the query sequence. The result of