Prefilter pVACseq and pVACsplice on a user-defined list of biotypes

[susannasiebert]

Currently, pVACseq does not prefilter transcripts on their biotype. We prioritize protein_coding transcripts when picking the best peptide and transcript in the aggregate report creation.

In pVACsplice we hard-pre-filter on protein_coding transcripts.

Instead we should define a new parameter --biotypes with a default of ['protein_coding'] that applies a prefilter on transcript biotypes both in pVACseq and pVACsplice. The aggregate report will continue to prioritize protein_coding transcripts when selecting the best peptide/transcripts. The docs should explain that "a few other, more speculative types (e.g. non_stop_decay and nonsense_mediated_decay) can give rise to neoantigens if users want to include them, they can specify their own list of biotypes".

[sufyazi]

hi I'm wondering when pVACsplice will be ready for use? I read your publication regarding pVACsplice but now I'm confused when I see that there is no pVACsplice in the pVACtools suite.

[susannasiebert]

Hi @sufyazi, we are currently doing extensive testing on this tool on various patient data and are hoping to release it before the end of the year. May I ask which publication you are referencing? While we've presented this tool at various conferences, we haven't published a full paper on it yet.

[sufyazi]

Hi Susanna, sorry if I was implying the wrong thing. I saw an abstract of pVacSplice here 10.1016/j.cancergen.2022.10.114 and wanted to find out more about it, only to realize it is not yet there in pVacTools. I just started using pVacTools at work for a new project we are embarking, and pVacSplice sounds perfect to incorporate into our pipeline, so I was just wondering.

Thank you for the update! I will look forward to it!

[susannasiebert]

@sufyazi pVACsplice is now available in the newest version of PVACtools (5.0.0).