Are you exclusively interested in testing gene pairs? Because if the genes are somehow functionally related it would make a lot of sense to test it as a “functional module” by applying the test to the whole group of nine genes or a subset that looks most promising to you. You may sort them beforehand (Edit->Rows->Sort by mutual exclusion).

If you want to make the 36 pairwise tests I think it is not absolutely necessary to perform multiple test correction, but it’s good practice – especially if you want to publish the result in a peer-reviewed article there is a good chance that you will be asked to apply a correction.

Lastly, yes by default the number of permutations is set to 10’000. If you wish to change that you may try to run the analysis via Analysis->Mutual exclusion & co-occurrence.

I hope I could help out. If you have more questions please feel free to add a further comment or use the mailing list.

Best,

Michael

If I have a set of tumor samples and a list of 9 genes with binary mutation information, to test the mutual exclusivity or co-occurrence of one of these genes with any of the 8 remaining genes I would have to test this one by one to determine the p-value for that test (eg. gene 1 vs. gene 2, gene 1 vs. gene 3, etc.). That would be 36 tests and 36 p-values of mutual exclusivity (and 36 for co-occurrence). Would you recommend correcting for multiple testing (0.05/36) or does the fact that each of these p-values has been determined through permutation testing negate the need for Bonferroni correction?

Also, am I right in thinking the number of permutation tests is 10,000 in each case?

Thanks,

Adam

]]>the minimum threshold for OncodriveCLUST is 3 (i.e. a gene needs to have at least 3 mutations to be included in the analysis). Unfortunately, in a data set of just 6 samples you have no statistical power for any cohort analysis.

best regards,

David

What would be the best threshold, minimum number of mutated samples a gene must have to be analysed by OncodriveCLUST, when you have a small sample size of 5 or 6 six samples?

]]>It is not yet decided if the course will be held this year. We will post the information here in the blog and/or twitter (@bbglab) when available.

Thanks for your interest in the course,

Nuria

]]>Due to my interest in this summer course, I would like to ask if you have any idea when it will be possible to apply for this summer course in 2014?

Kind regards.

]]>Thank you for your interest in Condel. The web server had some troubles a couple of weeks ago. This is probably why it kept crashing on you. It is now up and running.

Best,

Abel

]]>