CNV.output.tsv not generated

[Devin-D]

Hello working with several samples, all but 1 sample output the cnv.output.tsv. Have ran accucopy on this single sample many times, however no CNV output. Accucopy finishes with no error. Job finishes but the cnv.output file is not written. What are the reasons this woud happen?

[polyactis]

Can you send log files of the failed sample ? pyflow_log.txt pyflow_tasks_stderr_log.txt pyflow_tasks_stdout_log.txt

…

-- Yu S. Huang http://www.yfish.org/ https://sites.google.com/site/polyactis/

On Tue, May 16, 2023 at 10:49 PM Devin Dinwiddie ***@***.***> wrote: Hello working with several samples, all but 1 sample output the cnv.output.tsv. Have ran accucopy on this single sample many times, however no CNV output. Accucopy finishes with no error. Job finishes but the cnv.output file is not written. What are the reasons this woud happen? — Reply to this email directly, view it on GitHub <#23>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAF7C2JQLKVPP7NUAF3RWL3XGOHYJANCNFSM6AAAAAAYDYW4PA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Devin-D]

Hello thank you for getting back to me. In looking through the logs for this sample it seems the error is occurring in implementation of GADA
289 segments. 289 segments used. 181247 SNPs used. Calculating auto correlation ...Done. Calculating auto correlation shift-1 difference ... #mean is: -0.000277749, sigma is: 0.000149832 Done. Inferring candidate periods through GADA, run_type=1, left_x=-0.000315709, right_x=-0.00023979 ... Initiating GADA instance ...GADA done Found 0 candidate periods. ERROR: No candidate period discovered.

This stops the output of infer.out.tsv and cnv.output.tsv. I am not familiar with this error from GADA but have since run into the same error on 2 additional samples. These are 30X tumor normal samples and I have no reason to believe there is zero copy number variation, clonal or otherwise.

[polyactis]

if plot.tre.jpg is available , you can plot it here.

Although your sample is a 30X tumor-normal pair, our algorithm can fail in :

1. data too noisy, too many subclones, no periodic pattern could be inferred.
2. tumor purity too low (i.e. <5%). The tumor sample comprises mostly normal cells, which makes it similar to a normal sample.
3. too few copy number alterations.

In your case (289 segments, a bit low), my guess is either 2) or 3).