Phasing using SHAPEIT

Question

[genetic]

Hello,

I need to use SHAPEIT for phasing only since I will conduct CH (compound heterozygous) analysis for recessive rare variant. I will not perform imputation.

I am running SHAPEIT, and I see in the log file it says:

Parameters :
  * Seed : 1442251531
  * Parallelisation: 12 threads
  * Ref allele is NOT aligned on the reference genome
  * MCMC: 35 iterations [7 B + 1 runs of 8 P + 20 M]

I am still able to get *haps file for haplotypes for CH, however, I am not sure if I am doing correctly.

Is it ok to have "Ref allele is NOT aligned on the reference genome" notice on my log file?

I have one more question.

My input file is plink PED/MAP format, and on the SHAPEIT website (http://shapeit.fr/pages/m03_phasing/input.html), it says that SHAPEIT considers "0" as missing data.

And they suggested people to change the missing data character to "N" for example, use --missing-code options as follows:

shapeit --input-ped chr20.unphased.ped chr20.unphased.map -M chr20.gmap.gz --output-max chr20.phased --missing-code N

However, --missing-code N gives me an error ERROR: Non biallelic site pos=24118582 a=0

So, I did not use --missing-code N and run SHAPEIT:

shapeit --input-ped chr20.unphased.ped chr20.unphased.map -M chr20.gmap.gz --output-max chr20.phased

Would that be ok?

Thank you so much

[Endre Bakken Stovner]

It might mean that not all of your panel/reference alleles were used. This might be because your plink files are not all on the reference strand, see https://github.com/endrebak/snpflip for a solution (if this is the problem).

[Endre Bakken Stovner]

Your title is too general btw.

Answer 1

[Kevin Blighe]

For anybody else arriving here, I show a step-by-step that will help for pre-phasing with SHAPEIT2:

• C: Phasing with SHAPEIT

For phased imputation using the output of SHAPEIT2, take a look here:

• A: ERROR: You must specify a valid interval for imputation using the -int argument,

Kevin

[Shicheng Guo]

Is there anyway to perform haplotype phasing without imputation with Shapeit4? or imputation is necessary?

[Kevin Blighe]

I am not sure. I imagine that you can do just pre-phasing with SHAPEIT4

[summerday1112]

I just want to phase the vcf file and do not want shapeit impute the misssing genotype ./., do you know how can i do it