Background Variant interpretation is dependent on transcript annotation and remains time consuming and challenging. There are major obstacles for historical data reuse and for interpretation of new variants. First, both RefSeq and Ensembl/GENCODE produce transcript sets in common use, but there is currently no easy way to translate between the two. Second, the resources often used for variant interpretation (e.g. ClinVar, gnomAD, UniProt) do not use the same transcript set, nor default transcript or protein sequence. Method Ensembl ran a survey in 2018 to sample attitudes to choosing one default transcript per locus, and to gather data on reference sequences used by the scientific community. This was publicised on the Ensembl and UCSC genome browsers, by email and on social media. Results The survey had 788 responses from 32 different countries, the results of which we report here. Conclusions We present our roadmap to create an effective default set of transcripts for resources, and for reporting interpretation of clinical variants.
After decades of avoiding the demand to highlight one transcript per locus in Ensembl, we ran a survey 2018 to assay opinions across the scientific community. Ignoring the problem of ‘one transcript’ was not making the issue go away; many important genomic resources had instead adopted their own methods of selecting one transcript (e.g. HGMD, Ensembl, gnomAD, UniProt, ClinVar, etc.). Here we report our results and roadmap to create an effective default set of transcripts for resources, and for reporting interpretation of clinical variants.