sampleQC issue

[Xuan06]

Hi,Ted

when I want to use Atlas-CNV for my target sequence data, I get some result files. They are sample_id.cnv.FAILED_sampleQC_and_sampleANOVA or sample_id.cnv.FAILED_sampleQC. Do they mean that these results are failed? And don't I get successful CNV from my sequencing data? Can I change some parameters to get better results?

I re-post this issue in github issues, so other users can benefit too.
Thanks again for your help.

[theodorc]

Hi Xuan,

Thank you for your interest in using Atlas-CNV.

When you see a failed sample_id, it means it has failed the sampleQC=0.2 criteria (> 0.2). This was the valued used for the eMERGE panel of 109 genes with 45 samples per batch. Generally, if you use a similar experiment, this will ok.

However, if you have fewer samples or if your panel has more genes (and therefore more targets), then the sampleQC threshold should be adjusted slightly. Usually, I make a simple plot of all the sampleQC in a given batch and visually see if the values are consistent, and then you can change or set that sampleQC accordingly (see supplementary online materials of Atlas-CNV publication).

When you run the analysis, you should get a sampleQC outputfile (Sample_RPKM-means.... file). Look at the last column SampleQC and see if the values are consistently in the same range. Outliers would suggest true failed samples. But you should consider sample size too.

Thanks,
-Ted

[NTNguyen13]

@theodorc Hi, thank you for your answer. My samples are validated by Sanger, some of them contain multiple-exon deletion, when I check the RPKM matrix, I saw that they are lower than other sample. But those samples are reported as Failed Annova, and also not reported.

I also checked the Cscore and Pvalue matrix file, they also have significant lower values than others, but I don't understand the meaning of these 2 files.