hammerlab / paper-aocs-chemo-neoantigens

Manuscript on chemotherapy-induced neoantigens in samples from the Australian Ovarian Cancer Study

paper-201604

This repository has the notebooks, LaTeX source, data, figures, and supplemental files for the article "Chemotherapy weakly contributes to predicted neoantigen expression in ovarian cancer".

Running the analyses

Analyses were run in this order.

• 'samples' notebook
• scripts/annotate_variants.sh
• 'prepare and merge varlens annotated mutations' notebook
• 'neoantigen and mutation counts' notebook
• 'signature extractions with model' notebook
• 'prepare mutation counts for deconstructsigs' notebook
• R script: R/signature_deconstruction.R
• 'deconstructSigs results' notebook
• 'signature probabilities per mutation' notebook
• 'plot mutation and neoantigen signature sources' notebook
• 'plot extracted signatures'
• 'prepare additional files'

paper-aocs-chemo-neoantigens's Issues

Submit to BMC Cancer or JITC

address reviewer comments: clarify allelic fraction statement

Can you elaborate on why NACT treated tumors have fewer neoantigens expressed [line 52, page 8]. It may be my limitation as a clinician, but what do you mean by
"undetectable allelic fractions without the population bottleneck created by surgery"

address reviewer comments: why use cisplatin sigs when carbo is standard of care

I think that it would be helpful to address in the methods sections about why you utilized cisplatin exposed C. Elegans and G. Gallus [Page 4, Line 6] rather than carboplatin (the standard of care first-line chemotherapy agent of choice). I know that your discussion addresses this on page 8, Line 19, but it may help to introduce this earlier.

You also mention that the signatures for carboplatin may be different, would this change your assertion the increased neoantigen expression is from other sources (in place pre-therapy) rather than chemotherapy exposure?

address reviewer comments: room for more of an immunotherapy for ovarian bottom line in discussion?

address reviewer comments: add demographic table

Reviewer

I think from a clinical standpoint, it may help to have a "demographics" type table reflecting the information in additional file 1 to help the reader understand what type of samples and chemo exposures your cohort is representing.

Editor:

Please submit a new Table 1 that would be included in the printed manuscript, not just in an additional file.

address Bowtell's comments

Interested if anyone has a reference stronger than Brown et al for the claim that patients with highly mutated tumors who do NOT receive immunotherapy have better overall prognosis (CC @snyderca)

My TODO from David Bowtell's comments:

• Need to change conclusion to be that chemotherapy may be causing many subclonal mutations. We can’t conclude that a given cell has few chemotherapy mutations, only that those that do exist are below detection threshold.
• everyone got platinum but not everyone got a taxane. actually possibly 1 pt did not get platinum? need to compare from data from Liz.
• "Certainly, in CRC, ovarian and others, paients with a strong immune response have a better prognosis but is there data that more mutations is associated with better prognosis or just a greater chance of ICI response?" Need to respond: yes, there is such data (find it)
• "Just about all the ovarian immune checkpoint data so far relates to patients who have received prior Pt-based therapy….and still have had low response rates." This is a fair criticism, should maybe just mention it in the paper.
• Note that the C elegans / chicken studies were done in clonal organisms, whcih is different from cancer
• Correct for multiple testing in p-values
• Define xpf-1 etc knockouts early in paper