|August 19, 2016
Added a new Gene + ExAC Freq. sort option to variant search results which first sorts by Gene and then by ExAC allele frequency.
Gene Details dialog now includes a description of the gene's function from UniProt.
New 'Summary of Findings' link on the Project page presents a summary of family analysis statuses, tagged variants, and phenotypes entered.
Fixed bug where Variant search showed an error message when using very long gene lists.
|July 17, 2016
Variant Search results can be sorted by different criteria using the Sort By: drop-down that appears to the right of the Search button.
Gene Search now allows searching for variants in two or more projects at the same time.
Updated Gene Details tissue expression view to use GTEx release V6 which includes more tissues and samples.
On the project page, a table near the bottom now shows a summary of the number of Phenotips terms entered. This table is only shown for accounts with 'manager' permissions on a project.
PhenoTips™ has been integrated for structured entry of patient phenotypes.
For projects that were sequenced at the Broad Institute, we have access to the underlying sequencing read data (.bam files) and will now be able to show an IGV-like view of the reads directly in the browser. In projects with this feature enabled, read visualization can be accessed through a SHOW READS icon that appears next to each variant in the search results.
When a variant falls in a gene that has multiple transcript isoforms, seqr selects only one of these transcripts to show in variant search results. The choice of transcript may affect things like whether the variant is labeled missense or splice-disrupting, as well as 'HGVSp' and other annotations.
Before this update, seqr always chose the worst-affected transcript, but this sometimes led to obscure transcripts being shown for important disease genes.
To avoid this, we've updated the transcript selection logic to:
- ignore non-protein-coding transcripts (except when all transcripts for a given gene are non-protein-coding).
- of the remaining transcripts, select the worst-affected transcripts for the given variant.
- if there's a tie where multiple transcripts are worst-affected in the same way, see if any of them is canonical according to Gencode v19. If yes, use the canonical transcript. Otherwise, chose by alphabetical order of the transcript id.
For De Novo Dominant variant searches on families that have data for mother, father and child, more stringent filters are now applied as follows. A variant will only be shown if:
1) the read coverage in the child is no less than 10% of the total read coverage in the parents at the variant site
2) the variant GQ score is >= 20 in the child
3) the parents' variant call allele balance (number of reads supporting alt allele / total reads ) is less than 5%
- in this case, the setting from the adjustable Allele Balance slider is still applied, but only to the child, while the GQ slider value only applies to the parents.
Clicking on the variant effect (for example 'missense' in ) brings up a popup dialog with all transcripts listed. After this update, the popup dialog will now show which of the transcripts is canonical, and also the HGVSp for each.