Opened 7 weeks ago

Last modified 6 weeks ago

#1343 new task

Manual genotype check of standalone VCF files

Reported by: Nicklas Nordborg Owned by: Nicklas Nordborg
Priority: major Milestone: Reggie v4.34
Component: net.sf.basedb.reggie Keywords:
Cc:

Description

Sometimes it would be useful to be able to take a standalone VCF file and run it against the existing genotype VCF files to see if we get a match.

The current wizard uses the VCF file from a AlignedSequences item has all the functionality. The only difference is that we do not have related patient information for the standalone VCF file.

Change History (11)

comment:1 Changed 7 weeks ago by Nicklas Nordborg

In 6441:

References #1343: Manual genotype check of standalone VCF files

Started to implement the GUI for the wizard. It is possible to select files and submit to validation, but actual comparison against existing VCF files is not implemented.

comment:2 Changed 7 weeks ago by Nicklas Nordborg

In 6442:

References #1343: Manual genotype check of standalone VCF files

Started to implement the GUI for the wizard. It is possible to select files and submit to validation, but actual comparison against existing VCF files is not implemented.

comment:3 Changed 6 weeks ago by Nicklas Nordborg

In 6443:

References #1343: Manual genotype check of standalone VCF files

Implemented comparison. It seems to work well if used with other VCF files from the current genotyping implementation, but lacks a lot of error handling if used with VCF files that doensn't meet the requirements.

comment:4 Changed 6 weeks ago by Nicklas Nordborg

In 6444:

References #1343: Manual genotype check of standalone VCF files

Implemented functionality for loading a reference VCF to make it possible to compare VCF files on common grounds.

comment:5 Changed 6 weeks ago by Nicklas Nordborg

In 6445:

References #1343: Manual genotype check of standalone VCF files

Updated the VCF parser to be able to use both the ID column and the combination of CHROM, POS, REF, ALT columns for macthing.

comment:6 Changed 6 weeks ago by Nicklas Nordborg

In 6450:

References #1343: Manual genotype check of standalone VCF files

Fixed the "View genotypes" dialog so that it can work also when the VCF file is not linked to an alignment item. Some changes are related to working with VCF files generated via SNP array pipeline.

comment:7 Changed 6 weeks ago by Nicklas Nordborg

In 6451:

References #1343: Manual genotype check of standalone VCF files

Re-implemented the comparison to a two-phase analysis. The first step will find a patient that have a good match. The second step will then be used to output messages assuming that the VCF file is from that patient. It will also consider that case that there are multiple patients with a good match, but this should not really happen because then there is something wrong with existing alignments as well.

comment:8 Changed 6 weeks ago by Nicklas Nordborg

In 6452:

References #1343: Manual genotype check of standalone VCF files

Do not warn about low GQ score if the score is a "GenCall?" score.

Do not skip all comparisons just because the number of common genotypes is less than the ordinary threshold since it is usually possible to find the correct patient also with a lower number of genotypes.

comment:9 Changed 6 weeks ago by Nicklas Nordborg

In 6454:

References #1343: Manual genotype check of standalone VCF files

Added the VCF reference file we use for genotyping.

comment:10 Changed 6 weeks ago by Nicklas Nordborg

In 6455:

References #1343: Manual genotype check of standalone VCF files

Added a check for correct REF and ALT alleles if a SNP is found via ID, and implemented support for swapping the REF and ALT alleles to make comparisons available. If the alleles are swapped we also need to swap:

  • Genotype: 0/0 swapped with 1/1 and vice versa
  • AD: Change order of the values
  • AF: 1-AF

comment:11 Changed 6 weeks ago by Nicklas Nordborg

In 6456:

References #1343: Manual genotype check of standalone VCF files

Added a link to the "Case summary".

Since there is no information to register in this wizard, the "Register" and "Cancel" buttons were replaced with "Restart" as the only button.

Note: See TracTickets for help on using tickets.