(updated in WBPS13 to reflect archiving functionality and mention gene IDs in GFF dumps)
As sequencing projects progress and yield improved data, we occasionally update genomes on the site to provide the best resources available. An example of this is Schistosoma mansoni, released in WBPS11. The authors of the new S. mansoni annotation preserved identifiers where genes were unchanged, and otherwise assigned new identifiers. They kept a list of changes – updates, splits and merges – between the two versions, available here in full.
Annotation authors do not always try to preserve gene identifiers, and e.g. in our other recent updates of H. contortus and F. hepatica, the identifiers are all new. In such case, we run a pipeline that tries to reconcile the two versions, and while the results are not even close in quality to a manually curated list of updates, it is still frequently helpful.
This list of gene updates is then integrated into WormBase ParaSite. Searching for an old gene should bring you to an archived ID page, showing the protein sequence of that gene and a link to a new ID if there is one. The data is also available in the GFFs, under previous_gene_id field in the column 9 of the gene.
The list is as complete as can be given the methods used to reach it – either authors tracking gene updates, or our method based on sequence similarity and homology – and sometimes our users find themselves with an old gene mentioned in a publication etc. that isn’t connected to a new identifier. We would like to describe a number of strategies that can be used to reach desired content in such cases, and go through a worked example.
Our search facility is best at suggesting and retrieving our gene identifiers e.g. WBGene00001135 and gene families e.g. hedgehog but will also retrieve a number of different identifiers from e.g. UniProt or GenBank. Do try it first, and let us know if the search is not returning something you would expect it to bring back.
FTP dumps from previous releases
Our Downloads section contains a link to our FTP site, with data for all previous releases of WormBase ParaSite. If all you require is e.g. a protein sequence for an old identifier, you are done – otherwise, you can use the sequence to search further.
Given any sequence, you can retrieve a corresponding identifier by BLAST – perfect or near-perfect alignment across a large substring is almost always the right hit. Of course you might also find that your gene of interest has multiple copies, has been split in two in the new annotation, etc.
Other websites 🙂
Going through the UniProt search frequently gives good results when searching for gene families, and you might want to try the BLAST service at ENA or UniProt as an alternative to our BLAST.
We aim for you to able to retrieve all the data you need from WormBase ParaSite, but using multiple resources is very pragmatic and frequently optimal, and thus with every release, we reconcile our identifiers with a number of external sources. This process lets our gene pages link to corresponding pages in another resources in the External references section. To find a relevant WormBase ParaSite page given e.g. an INSDC or UniProt ID, you can use the previously described search facility.
Exonerate is a very fast tool for sequence alignment. It could come in handy if you need to achieve similar results to our online BLAST for a few hundred anonymous sequences. We hope you will not have to resort to it!
Chalmers et al (2015) have identified ten tegumental surface proteins in S.mansoni as important in understanding the host immune response to adult schistosomes. Two of the ten genes – Smp_081920 and Smp_158960 – are absent from the new annotation.
Smp_158960 turns up in search results under a different identifier – it is now Smp_345020. Looking in the External references table reveals that Smp_158960 is the only previous identifier for Smp_345020, and checking view all locations link reveals that Smp_345020 is the only gene annotated with Smp_158960 as a previous identifier, so the mapping is one to one. We can guess that authors modified the gene structure and decided to give the gene a new identifier.
Searching Smp_081920 doesn’t return anything in search, and it took me a bit of time to figure out what happened.
I knew I could get the sequence from the FTP site, but it saved me some effort to discover that the identifier is still hanging on in UniProt, which gave me the protein sequence.
I have then used BLASTP, which gave me a full match with Smp_166350 on about 100bp: long enough for me to potentially trust but perhaps omitted by authors reconciling old and new identifiers.
I got certain I am looking at the same gene after comparing the exon structure of Smp_166350 with the structure of Smp_081920, conveniently still online at our friends’ place, Ensembl Metazoa. The two are very similar, except Smp_166350 has three additional non-coding exons.
I was skeptical of these non-coding exons and wanted to see some expression evidence for them.
I opened the JBrowse display in ParaSite and enabled some tracks across a few studies and developmental stages, and my skepticism grew. I saw high peaks in three non-coding exons and no reads aligned to what is claimed to be the coding region in tracks for cercariae and juveniles. Then, in tracks for adult worms but also in a track for miracidia, I saw the opposite: reads in the coding region, and no expression for that UTR.
My conclusion from this data would be that the gene Smp_081920 is now Smp_166350, but the structure of Smp_166350 is not completely correct: the UTRs are spurious, and they probably make another gene instead. I have submitted this conclusion to the annotation authors – it will be corrected in future versions of the annotation.