Announcing WormBase ParaSite 12

We are pleased to announce the 12th release of WormBase ParaSite, bringing new and updated genomes, and better handling of old identifiers and history.

New and updated genomes

The biggest updates of the release are probably two tapeworm genomes: Hymenolepis microstoma (PRJEB124), an update to a chromosome-level assembly, and a new genome Taenia multiceps (PRJNA307624).

There are also new clade IV nematode genomes: root-knot nematodes Meloidogyne graminicola (PRJNA411966) and Meloidogyne arenaria (PRJNA438575) (alternative assembly), and a bacteria-feeding Acrobeloides nanus (PRJEB26554).

The rest of the updates are:

Data and tools

We have re-ran our comparative genomics pipeline, constructing gene trees and finding orthologs and homologs. We have also reran newest InterProScan (5.30-69.0) and our cross-references pipeline for all our genomes.

We re-imported all public RNASeq data from our collaborators, and did a round of minor improvements to the displays. We now use information from PubMed to let you find data sets corresponding to a publication of interest.

Archived gene IDs

New sequencing technologies let labs construct better genome assemblies, bringing access to chromosome level assembly data even to relatively small research communities. We are excited to see this trend. Each genome update brings new evidence and potentially unlocks research into previously forbiddingly difficult biological questions.

At the same time, insights gathered in work published using previous assemblies should stay accessible to the community, so there is a need to connect different assemblies with each other. As of this release, WormBase ParaSite will keep track of previous identifier versions at gene level and display annotation history. Authors of a genome update do not always provide a mapping to previous version, so we developed a pipeline to match up identifiers between genome versions.

Overview of new functionality

For an overview of how this now works consider Smp_340760, a Schistosoma mansoni gene. The gene model was revised in the past, twice: it used to be called Smp_044010, but in “Schisto_7.1” version of the annotation that we published in WBPS11 the authors changed gene structure enough that they decided to assign it a new identifier, and in “Schisto_7.2” which we publish now the gene model was corrected slightly.

Searching by Smp_044010 now leads to a page explaining that the identifier was deprecated and redirecting to Smp_340760. Over there, the history is represented by a diagram:

SchistosomamansoniPRJEA36577_Smp_340760

The site also displays previous protein sequences of transcripts, to help you carry forward any conclusions based on the previous gene model – the less the sequence has changed, the more similar results will be for e.g. BLAST matches.

ID mapping pipeline

We used authors’ mappings between annotation versions for updates of Schistosoma mansoni and Hymenolepis microstoma. Everywhere else we used an automated mapping pipeline, adapted from Ensembl gene build.

Pipeline description

The pipeline runs a sequence matching tool exonerate, scoring matches of exons between the two assemblies and propagating the scores onto the
transcript and gene level. The scores are then adjusted based on synteny – if a gene A is near gene B in the previous genome, A is mapped to A’ in the new genome, and there is a gene B’ near A’, the match of B to B’ is strengthened. Finally, best matches are iteratively taken out of the scoring, producing a list of pairs.

Results and benchmarking

We find the pipeline to be quite conservative even after we relaxed a few parameters around minimal match scores and similar values.  Typically only between a third and two thirds of the genes in the updated genome have a related past identifier:

Genome previous genes total previous version mapped new genes total fraction mapped
Ancylostoma ceylanicum PRJNA72583 11783 WBPS11 7564 15892 0.476
Ascaris suum PRJNA62057 17974 WBPS9 9468 15260 0.620
Fasciola hepatica PRJEB25283 16806 WBPS10 7564 22676 0.334
Haemonchus contortus PRJEB506 19430 WBPS10 11439 21869 0.523
Meloidogyne incognita PRJEB8714 45351 WBPS10 11977 19212 0.623

We also ran the automated pipeline on the S. mansoni WBPS10->WBPS11 update, comparing the results to a manual mapping obtained by annotators tracking individual identifiers. Our pipeline carried forward 5165 genes that authors considered to have none or minor changes, and 1347 genes with larger changes, onto an annotation with 10172 genes. The pipeline missed 2584 genes present somewhere in the manual mapping that were lost in the automatic one. It disagreed with the manual mapping in only 199 cases: some were genuinely wrong calls, and some are some were on par with manual mapping by being e.g. a mapping to a paralog gene.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s