Hi,

I'm using DNAcopy to segment copy number log-ratios, and I was wondering if somebody here is aware of reasonably well-tested code that jointly uses b-allele frequencies of heterozygous germline SNPs to achieve better results. 

It would be fairly trivial for example to find optimal log-ratio segmentation parameters (alpha etc.) by minimizing the segmentation error of mirrored b-allele frequencies (1-frequency if frequency > 0.5).

Or one could start with very liberal alpha values for the log-ratio segmentation, and then iteratively join neighboring segments when their b-allele frequencies are similar. 

I understand that there are existing complete software packages like ASCAT out there, but I would prefer using a simple wrapper function around DNAcopy's segment function if possible.

Any feedback is highly appreciated.

Cheers,

Markus

The segmentate function in the beadarraySNP package can use LAIR (a B-allele derivative) to improve segmentation. It actually segmentates the copy number signal first, and splits segments if subsegmenst are found with differing B-alleles