Meeting 17: Biweekly GIAB SV Integration Call Monday April 2, 2018 Purpose On the April 2 call, Lesley Chapman gave a demonstration of SVcurator a web based interface for the manual curation of structural variants. Justin Zook also gave a brief update on using results from Nancy Hansen’s assembly analyses to define high-confidence regions for SVs. Key Points ● Slides can be found here: ○ SVcurator App ○ Link to SVcurator App ● SVcurator app ○ App will allow users to manually curate 1000 insertions and 1000 deletions randomly sampled by size from the Ashkenazim Jewish Trio son ○ App will be made publically available by May 8, 2018 at Biology of Genomes ● Update: Proposed process for using assemblies to develop High Confidence BED Files ○ Only 355 extra calls found of the 13707 “Tier 1” calls within 1kb of extra calls in >2 assemblies ○ ~11000 out of 13000 Tier 1 calls are covered by both haplotypes in the PacBio/10x diploid assembly from Mark Chaisson, which is promising ■ Next steps include using results from all assemblies What’s Next ● Adam English from Spiral Genetics will give a report on using Tier 1 calls as a benchmark on our next call at ​3pm EDT [12pm PDT] ​on​ April 16 2018 Detailed Meeting Notes SVcurator App ● SVcurator is a platform that will allow users to manually curate 1000 insertions and 1000 deletions from the Ashkenazim Jewish Trio son ● Data generated by the manual curation app will be used as labeled data for machine learning classifiers ○ Possible labels include: ■ Genotype ■ Genotype Certainty/Confidence score ● Images displayed in the app: ○ svviz2 images that show alignments of reads from Illumina, PacBio, and 10x to structural variants ○ svviz2 dotplots ○ IGV images ● 1000 Deletions and 1000 Insertion site selection:

○

● ●

Randomly sampled from the v0.5.0 callset based on size where each of the following size ranges are represented: ■ 20-50bp ■ 50-100bp ■ 100-300bp ■ 300-400bp ■ 400-1000bp ■ 1000-5999bp ■ 6000bp+ Each user will curate a random set of SV images Goal: make the app publicly available by May 8 - The Biology of Genomes