Vcf filtering #31
Vcf filtering #31
Conversation
jkbonfield
force-pushed
the
vcf_filtering
branch
from
f19cc1d
to
ec4c580
Compare
There was a problem hiding this comment.
images/150x_DPQ30_lhist.png appears to be empty (and possibly not referenced)? Is it supposed to be there?
Rather than have workflow-cram.md etc, would it be possible to put the workflows into their own directory (so workflow/cram.md), so the root directory doesn't get too crowded?
workflow-filter.md
Outdated
|
|
||
| Additionally `bcftools call` has some options which govern output of variants. | ||
|
|
||
| Option | Description |
There was a problem hiding this comment.
Are all these non-breaking spaces necessary?
There was a problem hiding this comment.
Yes or it looks totally rubbish, but I've found a better solution in <img width=150/>.
There was a problem hiding this comment.
Also removed the DPQ30 images, added the ones I missed (BQBZ), fixed the rogue quote and done some shuffling to a subdirectory. Testing it atm. Thanks for feedback
workflow-filter.md
Outdated
| The most obvious filter parameter however is the QUAL field. | ||
|
|
||
| Bcftools can filter-in or filter-out using options `-i` and `-e` | ||
| respectively on the `bcftools view` or `bcftools filter1 commands. For |
There was a problem hiding this comment.
Missing closing back-tick here.
TODO: indels
Each has a table of contents linking to the others for ease of navigation. Maybe this isn't needed though as the menus are there also?
jkbonfield
force-pushed
the
vcf_filtering
branch
from
09cbea1
to
24a6ff2
Compare
- Moved workflows to their own subdirectory - Added BQBZ plots - Removed DPQ30 plots - Fixed table widths in a more elegant fashion - Fixed missing quote
jkbonfield
force-pushed
the
vcf_filtering
branch
from
24a6ff2
to
5edde64
Compare
|
After a ridiculous amount of trial and error to prevent 0 headings or 2 headings per page, I finally wrestled it into submission and I think this is now ready for review again. |
The workflows section was getting long, so I split it into separate documents.
I then added a VCF filtering guide, with images showing true vs false variants plotted against a range of INFO and FORMAT metrics. It's clear some of these can work well for filtering variants, although how to do so well is tricky as the values needed are often depth specific. The main aim is to teach people how to explore filtering on their own data sets (assuming they have a known truth set). Possibly we should improve on this by providing some scripts, rather than the one-liners embedded within the tutorial, although that is at least educational.
For now this is a draft PR as the parameters filtered on are tied to the changes in samtools/bcftools#1363, especially for indel filtering.
I think only once that bcftools goes live (not this release, but hopefully the one after) can we consider publishing a filtering guide.
A preformatted version of this PR (at the time of writing at least) can be seen here:
https://jkbonfield.github.io/www.htslib.org/workflow-filter/