compbio-UofT / medsavant

MedSavant is a search engine for genetic variants
22 stars 9 forks source link

Annotation performance needs tweaking #285

Closed jvlasblom closed 10 years ago

jvlasblom commented 10 years ago

Need to test with profiler and optimize annotator for all sizes of VCFs. (Small VCFs with variants widely spaced can be slow to annotate depending on how parameters are set).

jvlasblom commented 10 years ago

Another way to improve performance on batches of VCFs would be to process all the VCFs at once instead of serially, so that within a certain region on the chromosome there will be many densely packed variants from the various VCFs. By increasing the density in this way, we will require fewer tabix seeks.