segments with deletions become 'red' colored

[hw538]

Hi ichorCNA team,

not sure why the delections become 'red' in our genome wide plots.

if I understand correctly, all the deletions should be 'green' colored. would you have any hints?

here is an example attached. SLX-18760_SXTHSH03.FS_1.sort_genomeWide.pdf

[hw538]

in the attached pdf above, the chr 6, 9, 10, 13 have those 'red' deletions, quite strange...

[gavinha]

Hi @hw538

Label switching has occurred. Very likely this is a suboptimal solution and should not be selected as the final result.

I would recommend that you turn on the subclonal setting (--estimateSubclones TRUE and --scStates "c(1,3)") and or run for multiple restarts (solutions) for --normal and --ploidy.

Hope this helps.

Best, Gavin

[hw538]

Thank you gavinha, for --estimateSubclones TRUE, I didn't find this option in the wiki page, would you mean --estimateScPrevalence TRUE instead?

Best, Haichao

[hw538]

Thanks, by adding --estimateScPrevalence TRUE and --scStates "c(1,3)", the color label now becomes normal. and it is true that the light green seg indicating subclonal CNVs. see attached:

rerun_SLX-18760_SXTHSH03.FS_1.sort_plots_genomeWide.pdf

could I know further why a suboptimal solution would alter the color label?

(would the explanation be:

if I don't turn on the --evaluateScPrevalence option, the ichorCNA algorithm won't consider subclones, and finally, the algorithm thought it had chosen an 'optimal' path, but this one CANNOT reflect the true CNV scenario in the genome. i.e. the 'optimal' path treats those regions as 'amplification' (red-colored), but in reality, these regions are 'deletions' (green-colored).

Best, Haichao