Eric Talevich

If the gene has only 2 exons, then that may not be enough data points for the segmentation methods to pick it up as a distinct region. At that point...

It depends on the type of sequencing you're doing for these ctDNA samples. Are you capturing/amplifying a gene panel, enriching the whole exome, or sequencing the whole genome? You can...

Just use `--normal` once, followed by all the BAMs you want to treat as normal: ``` cnvkit.py batch target.bam --normal n1.bam n2.bam n3.bam --targets related_bed_file.bed ... ```

It can be helpful to plot each of these columns and look for outliers visually. In your case I'd recommend opening this table in a spreadsheet, sort by each of...

That is not expected. It's also surprising that each of those segments has only 1 probe on it. Could you share the corresponding lines of the input .cns file?

Thanks for the details. Since the segment has 0 weight I suppose it makes sense for you to add a post-processing step to remove 0-weight segments; it's not clear to...

No, it's not an option yet. CNVkit needs another step to estimate integer copy numbers; initially it only estimates read depth log2-ratios. https://cnvkit.readthedocs.io/en/stable/calling.html

It's the p-value resulting from a 1-sample t-test of that segment's probes's log2 values versus 0.0. https://github.com/etal/cnvkit/blob/master/cnvlib/segmetrics.py#L23 I agree it ought to be documented.

It sounds like your tumor samples are too pure for the assumptions that go with bulk tissue samples; I think you need to turn off `--drop-low-coverage` everywhere.

You might also do well with the `bintest` command to supplement your segmentation results. This will give you exon-level calls, which can be noisy for whole exomes, but in this...