loh file

[ph3344]

Hi jannatjiaz, I am so interested in your ChromoDethripsis methods. I am curious about one issue:

python CTResolveLowDensity.py --chromosome --bam <path to haplotagged bamfile (1) > --readnames <path to readnames list (2) > --haplotype1_reads <readnames pos chr haplotype1 (3) > --haplotype2_reads <readnames pos chr haplotype1 (4) > --haplotype_blocks <path to the haplotype blocks produced by whatshap stats (5) > --phased_haplotype_blocks <path to the phased haplotype blocks produced by code in haplotype phasing directory (6)> --loh_regions <path to loh file (7) > --outputdir \

The script need a loh file , and how can I process it, using which software and code. Thanks so much.

[jannatjiaz]

Hi! Thanks for your interest in the method.

The loh file is just a file with three columns: chr, start, end. It can be produced using any copy number calling software which will give you the copy number across all chromosomes in your sample. Extract the regions that are copy number 1 (or the appropriate number based on the ploidy of your sample) and create a file with just the locations. These regions are assigned to the wild-type allele in the code.

Hope this helps!

[ucsfpan]

Thanks so much for your help. I will follow your suggestion and use the method you described to create the LOH file. Your explanation was very clear and helpful. One more thing, I only have tumor samples, and it seems to be difficult to analyze because tools like FACETS usually require a normal sample as a control. Could you provide any guidance on how I can generate the LOH file in this scenario? Thanks again for your assistance. (ph3344 is my old account)

[jannatjiaz]

This code was written for samples with both tumour and normal data so I cannot provide first hand advice about copy number calling without a matched normal. That said, there are tools that exist that can do this on short read data which may be of use to you (for example ABSOLUTE). As long as you have loh regions in the correct format, these other algorithms can be used.

[ucsfpan]

Thank you for your valuable advice and information regarding the LOH tool. I will look into ABSOLUTE and ensure output LOH regions are in the correct format.