Purge faulty samples from the dataset

[fpbarthel]

Several patients are consistently failing M2 calling. Of note, as you can see in the table below, these patients are also blocklisted to varying degrees.

------------------------------------------------------------------------------------------------------------
M2-error log                                 error       blocklist status
------------------------------------------------------------------------------------------------------------
logs/mutect2/filtermutect/GLSS-MD-LP01.log   log10p      coverage excl in WGS *SS*
logs/mutect2/filtermutect/GLSS-MD-LP02.log   log10p      coverage excl in whole genome *SS*
logs/mutect2/filtermutect/GLSS-MD-LP03.log   log10p      fingerprint excl, coverage exclusion in WGS + WXS
logs/mutect2/filtermutect/GLSS-MD-LP04.log   log10p      coverage excl in WGS *SS*
logs/mutect2/filtermutect/GLSS-MD-LP07.log   log10p      coverage excl in both WGS + WXS
logs/mutect2/filtermutect/GLSS-MD-LP08.log   log10p      coverage excl in WGS, no R in WXS
logs/mutect2/filtermutect/GLSS-MD-LP10.log   errorRate   coverage excl in WGS *SS*
logs/mutect2/filtermutect/GLSS-MD-0019.log   errorRate   fingerprint excl
logs/mutect2/filtermutect/GLSS-MD-0084.log   errorRate   fingerprint
logs/mutect2/filtermutect/GLSS-MD-0085.log   errorRate   fingerprint excl
logs/mutect2/filtermutect/GLSS-MD-0137.log   log10p      coverage excl in both WGS + WXS
------------------------------------------------------------------------------------------------------------
coverage excl = samples excluded due to low coverage
fingerprint excl = patient excluded entirely due to mismatches

For many of the above patients the quality is so poor that the data is not usuable. However, for 4/11 patients the patient may be salvageable by dropping some of the faulty data.

@roelverhaak please sign off on the to-do list below

To-Do:

• [x] Drop seven patients from the dataset entirely
• [x] For the four patients marked as SS, drop only the faulty aliquots

We chose to leave the data for now and instead mark as deprecated so they will not be used in any analyses.

[fpbarthel]

Just to confirm in case you were wondering @roelverhaak, none of the 7 patients to be dropped were used in the paper, whereas the 4 to be kept were.

[fpbarthel]

Also LP03 failed CNV.

CNV-error log                                                             error          blocklist status
logs/cnv/denoisereadcounts/GLSS-MD-LP03-R1-01D-WGS-4NCFXD.allosomal.log   non-neg median   fingerprint excl
logs/cnv/denoisereadcounts/GLSS-MD-LP03-R1-01D-WXS-UWJG6Q.allosomal.log   non-neg median   fingerprint excl
logs/cnv/denoisereadcounts/GLSS-MD-LP03-R1-01D-WXS-UWJG6Q.autosomal.log   non-neg median   fingerprint excl
logs/cnv/denoisereadcounts/GLSS-MD-LP03-R1-01D-WGS-4NCFXD.autosomal.log   non-neg median   fingerprint excl

[fpbarthel]

Seven samples had all aliquots marked depricated

UPDATE analysis.files
SET file_format = 'aligned BAM (depricated)'
FROM (SELECT al.aliquot_barcode
    FROM biospecimen.samples s
    RIGHT JOIN biospecimen.aliquots al ON s.sample_barcode = al.sample_barcode
    RIGHT JOIN analysis.files f ON f.aliquot_barcode = al.aliquot_barcode
    WHERE case_barcode IN ('GLSS-MD-LP03', 'GLSS-MD-LP07', 'GLSS-MD-LP08', 'GLSS-MD-0019', 'GLSS-MD-0084', 'GLSS-MD-0085', 'GLSS-MD-0137') AND file_format = 'aligned BAM') subq
WHERE files.aliquot_barcode = subq.aliquot_barcode AND file_format = 'aligned BAM'

Four samples had WGS bams (but not WXS) marked depricated

UPDATE analysis.files
SET file_format = 'aligned BAM (depricated)'
FROM (SELECT al.aliquot_barcode, file_format
    FROM biospecimen.samples s
    RIGHT JOIN biospecimen.aliquots al ON s.sample_barcode = al.sample_barcode
    RIGHT JOIN analysis.files f ON f.aliquot_barcode = al.aliquot_barcode
    WHERE case_barcode IN ('GLSS-MD-LP01','GLSS-MD-LP02','GLSS-MD-LP04', 'GLSS-MD-LP10') AND file_format = 'aligned BAM' AND aliquot_analysis_type = 'WGS') subq
WHERE files.aliquot_barcode = subq.aliquot_barcode AND file_format = 'aligned BAM'

Update: made a spelling correction, depricated > deprecated

[fpbarthel]

Relevant aliquots marked as deprecated