I want to analyze the Cisplatin sensitivity of Top1mtKO. I choose the Drug sensitivity dose level parameter and Top1mt mutation expression. Data looks confusing to me. If the y-axis corresponds to viability, then data suggests non-mutants and mutants show similar responses to cisplatin. Could you please explain whether my explanation is correct? To check drug sensitivity, do you recommend other methods/parameters?
Volcano plot seems a bit confusing to me too. If you could explain, it would be great.
First, I’ll need to understand the question that you’re trying to ask of the DepMap data.
Your question sounds like you’re trying to ask something like, “Is the sensitivity to cisplatin different in the presence of TOP1MT knock-out” Is that correct?
The ideal way to ask that question would be to have data for the same models with and without TOP1MT knocked out, and then treated with cisplatin. Then you’d be able to make an apples to apples comparison where you could attribute any differences to the loss of TOP1MT.
However, the DepMap datasets do not include knockouts and treatment with a drug. The CRISPR datasets measure the effects on viability from knocking out genes. The drug datasets measure the effects on viability from treating with compounds.
To answer your question, it looks like you ran the correlation analysis asking “which drugs in the PRISM Repurposing data are correlated with mutation of TOP1MT”.
Using models which are marked as mutant TOP1MT is one way to approximate the question you want. However, I think you’re more likely to want to use the “Damaging mutation” dataset as opposed to the one you selected. The generic one-hot encoded “mutation” dataset has a 1 if there’s any mutation, not just damaging mutations. Many of those lines may be just silent mutations.
Looking at the mutation table for this gene, I see that we have ~44 models which have this gene mutated. TOP1MT DepMap Gene Summary About 1/3rd have damaging mutations, 1/3 are some protein change and 1/3 are silent.
There are about 14 lines with damaging mutations, which is some but not a lot. Regardless, we can take a look to see if there’s any obvious difference, taking what we see with a grain of salt due to the small sample size of the set.
I’d do this by going to data explorer and plotting the viability effect due to cisplatin treatment, and then “group by” the line’s TOP1MT mutational status.
Unfortunately, the sample size is worse than I said originally. I had said that there were 14 lines with damaging mutations, however, only two appear in this plot. This is likely due to us having mutation calls for 12 of those lines, but those lines not being included in the repurposing dataset. As a result, I don’t think we have enough data about TOP1MT mutants to say much conclusively about the effects it has on cisplatin.