Revisit the filtering strategy for the UMI workflow

[ivadym]

Need

A reduced number of variants are currently being uploaded to Scout for the Balsamic UMI workflow. Therefore, we need to improve the filtering in order to provide customers with a useful list of variants.

Suggested approach

• Revisit the strategy for filtering UMI cases and adapt the filters accordingly
• Consider uploading research variants after making filters less stringent

Requests/suggestions/bugs solved by the feature

• Reduced number of variants in Scout

Can be closed when

• Customers can use Scout to go through the SNVs

[ivadym]

Some of the latest UMI cases with their number of variants:  

Case 1 (downsampled to 40M):
- 53 raw variants
- 0 research variants
- 0 clinical variants
 
Case 2 (downsampled to 60M):
- 161 raw variants
- 20 research variants
- 19 clinical variants
 
Case 3 (downsampled to 80M):
- 287 raw variants
- 85 research variants
- 78 clinical variants
 
Case 4 (PALKTTR040):
- 3407 raw variants
- 167 research variants
- 126 clinical variants
 
Case 5 (PALKTTR040):
- 3132 raw variants
- 108 research variants
- 77 clinical variants
 
Case 6 (PANKTTR080):
- 404 raw variants
- 174 research variants
- 174 clinical variants
 
Case 7 (PANKTTR080):
- 543 raw variants
- 254 research variants
- 254 clinical variants

[mathiasbio]

This might be somewhat affected by the changes in: https://github.com/Clinical-Genomics/BALSAMIC/pull/1176 regarding the updates for the fastp rules, where quality trimming for the UMI-workflow will be removed. Previously quality trimming was done prior to extraction of the UMI sequences which could decrease the sizes of the UMI families, and fewer consensus reads in the final bam. Maybe we can check these numbers again after these fastp-changes 🤔

[ivadym]

Closing as there are already specific requests in: https://github.com/Clinical-Genomics/BALSAMIC/issues/1336