polyactis
unable to infer tumour purity
Hi there, Hope you're well! I am running accucopy on multiple paired tumour-normal WGS samples. It has worked perfectly on most of them, except for in one sample where the infer_out.txt file is empty. I have checked the infer.status.txt file and the last line shows this error: ### Best period from likelihood: 0 best_purity: -1 best_ploidy: -1 Q: -1 logL: 0 best_no_of_copy_nos_bf_1st_peak: 0 first_peak_int: 0 ERROR: logL 0<=0 or best_purity -1 <=0!
Please could you let me know what this means? I am reading it as it is returning the purity to be 0%, but this would be impossible.
Thanks very much
Accucopy will fail if the sample is too noisy or undergo a little copy number variations. Can you share the png file plot.tre.png in the result fold? This file will told me why Accucopy failed.
Accucopy will fail if the sample is too noisy or undergo a little copy number variations. Can you share the png file
plot.tre.pngin the result fold? This file will told me why Accucopy failed.
Hi, Thanks very much for youre reply.
Here is the plot shown.
In the TRE plot, one major peak represents one major event of copy number variations, and the peak near 1.0 represents copy number 2. There is one major peak in the figure, which means this sample underwent a little copy number variations. Maybe this sample comes from the early stage tumor. In this case, we can consider the methods only using SNVs data to get purity of tumor samples. Maybe you can try TPES, a method inferring tumor purity from SNVs (https://academic.oup.com/bioinformatics/article/35/21/4433/5490858)