Extract Simulated 23andMe (v3) Style File from Whole Genome BAM File

Thomas Krahn of YSEQ.net has developed a way to extract a simulated 23andMe (V3) style file from a Whole Genome BAM file. This is a different procedure than the one I mentioned in Convert 23andme V5 RAW to Gedmatch classic which focused on taking a v5 chip and converting it into a v3 format. I am not seeing it on YSEQ.net now, but they used to mention providing a simulate 23andMe raw data file if you ordered a Whole Genome Sequence (WGS) from them. Even if they no longer offer it, Thomas provides the information in the GitHub link above on how you can do it yourself.

A couple of key points made by Thomas:

SNP calling of huge BAM files requires more than average computing power, disk space and RAM. The minimum recommended setup is 4 Gbyte of RAM and 100 Gbyte free disk space. A 64 bit processor is recommended.

Of course you need a hg19 referenced, sorted and indexed WGS BAM file. I haven’t tried an exome BAM file, but it might work, since the 23andMe SNPs are often in exome regions. This is certainly sufficient for single SNP diagnostics, but it may be problematic for segment analysis or phasing.

A WGS .BAM file tends to be very large which is why you need so much unused space. Not mentioned, but you don’t want to do this on a machine where 100 GB would reduce your available hard drive space too much since you often run into problems when your unused hard drive space drops to less than 20% although some machines can handle it dropping to 10%, depending on a lot of factors. I wouldn’t recommend it on anything less than 500 GB or 1 TB hard drive and that’s if your unused hard drive space is fairly large.