- We propose a novel method for crowdsourcing a protein family database. We discuss how we intend to identify novel groupings of proteins from user sequence similarity search, and how text mining will be applied to assist in annotation of these novel groupings, and more broadly as an enrichment of protein sequence similarity search results. We intend to use entity linking to identify literature which discusses proteins found in the search results, and present those publications which are likely to be the most useful to curators and sequence similarity search users alongside the sequence search results.
Protein families are groupings of proteins which have an
evolutionary relationship. Pfam is a database of protein families,
which has been maintained since 1996 [
HMMER is also a valuable tool for discovering families. It
can perform sensitive sequence similarity searches against a
target protein sequence database, and is able to find homologous
proteins distantly related to the query sequence [
Since 2011, Pfam has used Wikipedia as the primary
repository for family annotation. If an article already exists,
curators link it to the family, and the contents of the article is
mirrored on the protein family's web page. For families which
do not have a Wikipedia article, users are encouraged to create
a new article [
Constructing a new Pfam entry requires that a curator searches the scientific literature for evidence of the existence of the new protein family, and for the possible function and structure of the members of the family. The curator’s aim is to identify literature which mentions a member of the family, and the most useful papers will be those which go into detail about the protein’s structure or function.
Identifying literature which mentions a particular protein is
more complex than searching the literature for the protein’s
name. Biologists may use several different names and
abbreviations for the same protein. Additionally, the same name
may be shared across several species [
Previously, PubServer combined protein sequence similarity
search using PSI-BLAST with literature retrieval [
Our aim is to introduce crowdsourcing to the construction of the families, and to facilitate easier annotation of families by the application of text mining.
HMMER can be used to identify new protein families. This could happen intentionally, when a Pfam curator makes additions to the database, but it can also occur incidentally, when a sequence similarity search user's query sequence happens to produce an HMM which matches against sequence regions which are as yet not matched by any existing Pfam family. It can also occur that a user's search matches a very high proportion of the sequences matched by an existing Pfam family, and further additional sequences which are likely to be homologous with the Pfam family. In this case, the HMM produced by the user's search is a potential improvement over the HMM which currently defines the family.
Presently, such potential improvements to Pfam will only be incorporated if the user identifies that they have found a grouping of proteins unknown to Pfam, and makes the effort to contact Pfam's helpdesk. We propose that users should be alerted to this situation and prompted to submit their search's sequence profile HMM to Pfam.
We believe that this has two advantages over an alternative system which would silently capture HMM's from sequence similarity search. Firstly, the user who identifies the new or improved family gets credit for their improvement, which we consider to be both an ethical necessity, and an opportunity for increasing community engagement. Secondly, we hypothesise that the user who performs a sequence similarity search which identifies a novel protein family is more likely to have the knowledge required to annotate that family with relevant literature or other metadata than Pfam curators are, resulting in higher quality annotation. This second advantage also applies over a potential alternative system which automatically performs sequence similarity search with random sequences which are not matched by any Pfam family. While this would likely produce many novel families, the curation workload to actually incorporate these families into Pfam would be high.
Once a user has been made aware that their sequence similarity search represents a potential improvement to Pfam, we would like to ease the process of adding annotations to their new family. In particular, we would like to highlight literature which is relevant to their sequence similarity search.
Presently, when researching a protein found by a sequence
similarity search, users could look at the literature linked to the
protein in a protein database such as UniProt. However, the
literature citations for an entry are not intended to be exhaustive,
but a representative sample [
We anticipate that literature search will be useful not only to curators and users constructing new protein families, but also to sequence similarity search users in general, as a tool for researching homologues of an uncharacterised protein sequence.
Our proposed method is to first use the BANNER named
entity recogniser, trained on the BioCreative II gene mention
data set, to identify all possible mentions of genes and proteins
within the open access subset of PubMed Central [
We intend to train and evaluate this method by using the references attached to each entry in the manually annotated subset of UniProt, SwissProt. Each publication in the reference list of a SwissProt entry has a ‘scope’, which describes the elements of the entry that the publication has been cited for. For example, an entry may have one reference with a scope of ‘nucleotide sequence’, which describes the sequencing of the gene which codes for the protein. The proposed system can be evaluated by its ability to extract the elements of the manually curated reference list from the literature, and rank the publications in order of the relevance which the manually annotated scopes imply.
Both of these components will form a new curation interface to the HMMER web service. This will be available for use by Pfam's curators, and by users who have identified potentially new or improved Pfam families through their protein sequence similarity search query.
We would like to thank the anonymous reviewers for their helpful comments.