Skip to content

Sprint meeting 2017 04 05

Jump to bottom
bjornnystedt edited this page Apr 6, 2017 · 8 revisions

Sprint meeting 2017 04 05 9.00-10.30, Uppsala BMC Navet, E10:3309 or video

Virtual meeting room: SciLifeLab@vconf.kth.se ; Pinkod: 8800; Rumsnummer: 990 SIP/H323: connect to SciLifeLab@vconf.kth.se. Web browser: https://vconf.kth.se/. Set SciLifeLab as conference room and ad your name below

Status update

The SV project has been running on low steam for a while for various reasons: limited staff, cluster downtime, focus on getting the SweFreq website in shape and prepareing the SweRef manuscript submission. Intentions have also shifted a bit, with Manta being more accepted as the default tool. We are now close to release of a first simple WF, coded in NextFlow, installed on Bianca. In its most easy form it is just a convenient way to run Manta, with the possibility to include a first rough filtering step based on INS/DEL/DUP/INV frequencies from the SweRef cohort (or other sources).

We could rather easily extend this as a way to run and collect output from multiple callers (FermiKit is already implemented). This output can then be used for more sophisticated downstream filtering. This is outside the scope of this project for now, but might be included later if consensus appear and there seems to be a community need for it. (See eg. Jesper et al, https://github.com/J35P312/FindSV, https://github.com/J35P312/SVDB).

SV cohort callers like GenomeSTRIP do not really fit in this framework (as it is sample-by-sample). These options are being investigated in related research projects, and we will await the outcome of these efforts before trying to provide any community WF for this.

Goals

First release version of the WGS-structvar WF.

Final(?) features to add in current sprint

  1. Add post-filter step för reciprocal overlap with bed-files (eg from SweRef freqs)
  2. Avoid stall if 1 sample fails; run all which works and report all fails at the end.

Future

  • Adding more callers? (TIDDIT? CNVnator?)

  • Close project and move to maintenance?

    Decision: Finish Phase 1 of the project, including release of the tool as is now (Manta + Fermi) and include in the SweRef article. In parallel, Daniel Nilsson et al will try to move FindSV into Bianca. Depending on success, we will either * use FindSV as the recommended user tool

  • implement the FindSV strategy into the wgs-structvar code (if this is better implemented for Bianca users), and continue FindSV potential further development in the wgs-structvar branch.

Goals for coming sprint

  1. Test post-filter step för reciprocal overlap with bed-files from SweRef freqs.
  2. Finalize sample fail handling; run all which works (and report all fails at the end if possible).
  3. Get an idea on if FindSV runs smoothly on Bianca.

Next meeting

Clone this wiki locally