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
Chromosomal arm addition is fundamentally incorrect #18
The example code relies on suppressing errors rather than doing the correct work when evaluating chromosomal arms, as it assumes there is at least one segment spanning the centromere, which is at least not the case for the QDNAseq example (I didn't test the controlFREEC data) and so it fails. This is silently suppressed, so the code goes on, but it only works by chance like this (and this is repeated for each chromosome, twice in the file for the first and the second optimization).
This was spotted when debugging the threshold finding code.
PS: I'm aware of the supposed "version 2" of this method, and the fact that the publication is out, but as there is no code available for shallowHRD 2 and thus only the laboratory which made it is able to run it, for all intents and purposes it does not exist at all. Any independent researcher / group wanting to use the algorithm will have to resort to this public version.
The example code relies on suppressing errors rather than doing the correct work when evaluating chromosomal arms, as it assumes there is at least one segment spanning the centromere, which is at least not the case for the QDNAseq example (I didn't test the controlFREEC data) and so it fails. This is silently suppressed, so the code goes on, but it only works by chance like this (and this is repeated for each chromosome, twice in the file for the first and the second optimization).
This was spotted when debugging the threshold finding code.
PS: I'm aware of the supposed "version 2" of this method, and the fact that the publication is out, but as there is no code available for shallowHRD 2 and thus only the laboratory which made it is able to run it, for all intents and purposes it does not exist at all. Any independent researcher / group wanting to use the algorithm will have to resort to this public version.