CYP2D6 haplotype reconstruction

Question

[khn]

Hello,

I am trying to reconsturct haplotype of CYP2D6, using PHASE2.1.1 after GWAS. I know that there are limitations to reconstruct for CYP2D6 just from SNP given deletions, etc. Is anybody familiar with the haplotype reconstruction for CYP2D6 by PHSE2.1.1?

1. I used the tag SNPs (based on the report by PharmGKB) for common star alleles in our ethnicity, but after reconstruction by PHASE2.1.1, the frequency of the haplotype did not match with previous reports.
2. I used the SNPs in the locations of CYP2D6 that were included when genotyping, but again, the frequency of the haplotype did not match with previous reports after using PHASE2.1.1.

Does anybody know how to resolve?

Thank you,

Answer 1

[raunakms]

Refer to this paper: Cypiripi https://academic.oup.com/bioinformatics/article-lookup/doi/10.1093/bioinformatics/btv232

[khn]

Thank you so much. I am trying to do haplotype analysis using Cypiripi, and can you/anyone please teach me how to create the following files in order to run Cypiripi as I have difficulties to move forward after reading the paper?

In order to run the tool, I need to use the code as below.

python cypiripi.py --fasta reference --fastq [mygenome.fastq] --cov [coverage]

So, I think I need to create the files as below.

1. fasta reference -- should I use the "reference.combined.align" which was packaged together? So I do not need to create something new?
2. fastq is your interleaved and paired .fastq file -- Is it possible to use the files that were used for GWAS?
3. cov -- So if my patients are 1500, can I just write as --cov 1500?
4. If my samples are 1500 or so, can I just start from "2" of "How to run Cypiripi on large samples"? During the process, could I make sam files?

Thank you so much for your help!

[inumanag]

Hi @khn,

author of Cypiripi here. Yes, for now FASTQs of WGS experiment are required (you can extract them from BAM, but then results can be non-optimal). On the other side, we have significantly improved in-house version which does CYP calling directly from BAM (not yet published). Let me know if you are interested.

As for your questions: - Yes, just use reference from the package - Sorry, I never worked with GWAS before, so no idea what would be its input? - Yes, but coverage we ask is coverage per chromosome. So, if you sequences your sample at 40x, you give Cypiripi 20. - 1500x is rather high for the prototype version of Cypiripi @ Github (it can run, but it will require a lot of memory). Again, you can contact me privately about the development version which can easily handle this size.

Cheers, Ibrahim

[khn]

Thank you so much in details.

Currently I do not have data of BAM or fastq, so I need to find out the different way.

But thank you so much!

[khn]

I think that we could use this tool only when we have fastq files. But thanks for your help!