search

aeeckhou / shallowHRD

This method uses shallow Whole Genome Sequencing (sWGS) and the segmentation of a genomic profile to assess the Homologous Recombination Deficiency of a tumor based on the number of Large-scale Genomic Alterations (LGAs).
30 stars 13 forks source link

Error or typo in the script #4

Closed zhqingit closed 2 years ago

zhqingit commented 2 years ago

Hi,

In script "shallowHRD_hg38.R", it seems you are using (quantile 75 -quantile 25)/2 (line 758) as the filter, but in the paper, it is (quantile 75 + quantile 25)/2. And I am not sure if 124000000 in line 453 is typo. And I wonder if you can write more details about the algorithm, it seems lots of actions after the threshold finding are missed in the paper. Thank you!

Best, Qing

aeeckhou commented 2 years ago

HI Qing,

Thank you for the message and sorry I took so long to answer I completly missed your comment. You are right, the typo is actually in the article, it is (quantile 75 -quantile 25)/2 ! For the line 453 it will not affect the results but one value should be kept indeed, I changed it in the code for the old version for hg38.

For the software, the CNA cut-off based on the diffrence between large segments of the genomic profiles helps to actually merge large then small segments to finally_ smooth the segmentation. If their difference in terms of value are below the CNA cut-off then two segments are considered to be on the same CNA level and are merged. The new version that I'm provided actually is improving those steps but also provide a more robust detection of CNA cut-off.

Thank you for your comment!

Best regards, Alexandre