Hi,

I am running DEsingle (https://pubmed.ncbi.nlm.nih.gov/29688277-desingle-for-detecting-three-types-of-differential-expression-in-single-cell-rna-seq-data/) to compare two clusters of cells identified by Smart-seq2 scRNA-sequencing. I have filtered the genes before running DEsingle to retain only genes expressed in at least 10% of cells from one of the two clusters (5379 genes retained). DEsingle outputs the results but I get the following warning message. How concerned should I be about this message and should I trust the results?

Warning message: In MLE(ll3, start = listeta.mu = eta.mu, eta.sigma = eta.sigma, : possible convergence problem: optim gave code=1 function evaluation limit reached without convergence (9)

My code is as follows (sce is a SingleCellExperiment object):

DEsingle_genes <- DEsingle(counts(sce), group = clusters)

Many thanks,

Lucy

Hi, Lucy,

Sorry to see this and reply late. Due to the complexity of ZINB model, for some specific data the parameters estimation could not reach the highest precision in R and thus will show the warning message. While it is the best ZINB parameters estimation results we could get now, and the precision is OK to use for the hypothesis test and DE genes detection. So it is OK to use the results.

Best, Zhun