Just a quick one this time.
The slides for the talk I gave at the GA4GH 6th Plenary meeting in Basel, on 3rd October 2018 are now online at https://www.slideshare.net/JamesBonfield/cram4
This talk covers CRAM v4.0 proposal, which could also be v3.1 if we restrict it to purely codec improvements.
Note that this hasn't actually changed much since the MPEG-G talk I gave 3 years ago! It's improved slightly, but back then it was a proposal, and right now it still is. In the interim there simply wasn't any strong community desire to eek out another 10-20% saving given it cost more CPU, but with NovaSeq we see we can gain at least 10% with no real CPU hit (more if we want to spend CPU of course) and there are other issues such as extra-long chromosomes that require format updates anyway.
No comments:
Post a Comment