While running DEseq for DGE, I am facing following problem. Error in preplot.locfit.raw(object, newdata, where, what, band) : ?? NA/NaN/Inf in foreign function call (arg 2) Calls: estimateDispersions ... predict.locfit -> preplot.locfit -> preplot.locfit.raw -> .C In addition: Warning messages: 1: In lfproc(x, y, weights = weights, cens = cens, base = base, geth = geth,? : ?? compparcomp: perfect fit 2: In lfproc(x, y, weights = weights, cens = cens, base = base, geth = geth,? : ?? procv: parameters out of bounds 3: In lfproc(x, y, weights = weights, cens = cens, base = base, geth = geth,? : ?? max_nr not converged Execution halted My file does not contain any "NA or empty cells", I have tried every possible thing but still not able to find the problem behind this error. Kindly suggest me some solution for this. I have attached the subset of input file along with this mail. Waiting for your positive response.] Regards, Nisha -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: Final_read_count_table.txt URL: <https: stat.ethz.ch="" pipermail="" bioconductor="" attachments="" 20140705="" 01a94730="" attachment.txt="">

Hi Nisha First the standard advices: 1. Please always post your complete code and the seesion info. Please do read the posting guide. 2. Please consider using DESeq2 rather the DESeq. 3. As you do not have replicates, you cannot perform a proper statistical analysis anyway. Now to your data: 4. Your read count table (apart from lacking replication) looks _very_ weird. Whenever a gene has more than 2 reads in one of the samples, it always has always _exactly_ the same count in both samples! This is a impossible to have happened by chance. Something has gone completely wrong in creating the table. So, it's no wonder that the dispersion fitting function gets confused. Simon