callVariant generates fusion derived peptides > 25mers when using --cleavage-rule None

[Geneece]

Hello, when running my fusion data through your no-cleavage branch, I noticed that my fusion derived peptides were larger than the default setting of 25mers.

I ran your test data (star_fusion.gvf) through and had the same issue.
I ran both your test data and my data through moPepGen with enzymatic cleavage (trypsin) and did not have the issue.

Any advice to address this output would be greatly appreciated, thank you.

[zhuchcn]

Currently, when --cleavage-rule None is used, the output peptides are no longer between max and min length. Instead, they are the superstrings of all variant peptides with up to X number of continuous amino acids without any variant. This means, if you have a variant every 10 amino acids, the peptide will just keep going.

So for fusion, because the entire C-terminus peptide sequence after the breakpoint is novel, it may just output the entire fusion protein. You can see the peptide table

#sequence	header	subsequence	start	end	feature_type	feature_id	ref_start	ref_end	start_offset	end_offset	variant
RQRDLRRALDTALRRVRGCPAQQAHSGGL	FUSION-ENST00000614167.2:1228-ENST00000642151.1:50|1	RQRDLRRALDT	0	11	transcript	ENST00000614167.2	968	1001	0	0	.
RQRDLRRALDTALRRVRGCPAQQAHSGGL	FUSION-ENST00000614167.2:1228-ENST00000642151.1:50|1	TALRRVRGCPAQQAHSGGL	10	29	transcript	ENST00000642151.1	48	105	1	0	FUSION-ENST00000614167.2:1228-ENST00000642151.1:50

So the amino acid 0-11 are from the left transcript, and amino acid 10-29 are from the right transcript.

This is definitely going to be a problem if you have fusion + SNP/SNV data, cuz it will hit the exponential complexity.

[Geneece]

I see, then non-specific cleavage cannot be used to generate peptides derived from fusions?

One additional question I have, does this branch output peptides that cannot be produced by the chosen canonical proteome database? This was described in your paper, however, I want to confirm if this is also the case when using non-specific cleavage.

Thank you for your time!

[zhuchcn]

I see, then non-specific cleavage cannot be used to generate peptides derived from fusions?

Sorry, I don't understand this question. The peptide RQRDLRRALDTALRRVRGCPAQQAHSGGL in my example is derived from fusion.

One additional question I have, does this branch output peptides that cannot be produced by the chosen canonical proteome database? This was described in your paper, however, I want to confirm if this is also the case when using non-specific cleavage.

Yeah, so --cleavage-rule None will not use the canonical proteome as a denylist currently. This is because right now, the length of output peptides is not in a set range, so it is hard to check whether a peptide of any length is a substring of any canonical protein. But someone else also suggested that we should instead output all possible variant 8-11 mers, using a sliding window. Once I have a chance to implement that, we will be able to remove canonical peptides.

The original paper was designated for shotgun proteomics. The immunopeptidomics and neoantigen support is novel and I unfortunately have limited bandwidth for this so it's moving slowly.

[Geneece]

Thank you for clarifying, this response answers my question. Yes a sliding window of all possible variants would be of interest for me as well.