Hi, I have been attempting to use DESeq2 as described in the manual to test for the significance of gene retention between groups of samples (multiple replicates). my design has both two categorical and 5 scaled continuous variables After running DESeq, I found a message saying("81 rows did not converge in beta, labelled in mcols(object)$betaConv. Use larger maxit argument with nbinomWaldTest").

Then, I run it like(dds <- nbinomWaldTest(dds, maxit=100) but still doesn't converge Attempting to increase the "maxit" number (10,000, and 1000,000 ) does not seem to help

What would be the solution?

Thanks for your help?

Check collinearity of the variables in your design. This can be an issue with convergence. Also, you can check the baseMean of the genes that do not converge, and possible remove these by low base mean or too few samples with a count >= 10.