Proper use of short_anchor

[cpsosa]

I am using -A to set the short_anchor but I do not see changes in the fusion supporting reads ( split_reads1 split_reads2 discordant_mates coverage1 coverage2 ).

Am I missing additional information?

Thanks

[suhrig]

Which values did you try? If you set it high enough, you should see fusions disappearing.

Arriba measures the longest distance from the fusion breakpoint to the furthest aligned base of a fusion-supporting read. If the distance is smaller than the given threshold, the fusion is discarded. This filter acts on the fusion level, not on the read level. It either discards the fusion as a whole or does nothing. You won't see changes in the read counts.

There is usually no need to tinker with this parameter. May I ask what is your motivation to change it?

[cpsosa]

-A 10
-A 30
-A 60

These are the values I tried around the default of 23 bp.
KIAA1549 BRAF -/- -/- chr7:138552721 chr7:140487384 CDS/splice-site CDS/splice-site duplication 11 10 3 2463 5806
KIAA1549 BRAF -/- -/- chr7:138552721 chr7:140487384 CDS/splice-site CDS/splice-site duplication 11 10 3 2463 5806
KIAA1549 BRAF -/- -/- chr7:138552721 chr7:140487384 CDS/splice-site CDS/splice-site duplication 11 10 3 2463 5806
KIAA1549 BRAF -/- -/- chr7:138552721 chr7:140487384 CDS/splice-site CDS/splice-site duplication 11 10 3 2463 5806 <----- DEFAULT

[cpsosa]

My motivation was to try to increase the number of supporting reads.

[cpsosa]

This particular fusion has a low number of supporting reads.

[suhrig]

I see. This parameter won't help you. Increasing the number of supporting reads is only possible when you see a value in the filters column with a number next to it in parentheses. These are all the reads that are visible to Arriba.

Otherwise, you need to optimize the STAR alignment parameters.

[suhrig]

If you like, I can offer my help to try to improve the parameters. I will need the FastQ files for this, though.

[cpsosa]

Thanks! This answers my question.