STITCH Set up - GBS-like data

[Pietro de Mello]

Hi there Dr. Davies et al.,

Dr. Davies, thanks for developing and making STITCH available. I'm excited to see how it performs with my data. I was able to install and run STITCH on the mock data provided in the STITCH github page, however I am not sure how to implement the program on my dataset. I'm afraid my questions might be obvious, and I apologize for that in advance.

Let's get to it:

First, a broader question:

Q1: In your 2017 webpage you mention that STITCH was applied to GBS data in mice, I was wondering how successful it was. There are few available imputation software options available, even less for GBS-like data, and I'm curious if STITCH would be able to give me reliable imputation.

Then, the technical questions.

Info:

    Data type: Intermediate depth (~ 10x) GBS data from non-model organism

    Genome available: Yes

Q2: At the beginning of your GitHub's README you say that STITCH "runs on a set of samples with sequencing reads in BAM format, as well as a list of positions to genotype, and outputs imputed genotypes in VCF format." However, in your toy example you give the program additional parameters, such as: "--chr=chr19 --bamlist=bamlist.txt --posfile=pos.txt --genfile=gen.txt ." My main question is: How did you produce your pos.txt and your gen.txt file?

Q3: Are the bamlist.txt and the gen.txt file mandatory for the program to run? Will the program look for the bam files in ./ if a bamlist is not given to it?

Q4: I noticed in a previous post that STITCH runs at one chromosome at a time. I also have a draft assembly that I'm working on. GBS data only makes matters worse, since loci are spread apart, which means that some scaffolds might only have a handful of reads. Am I right to assume that STITCH is 'blind' to other chromosomes, and that imputations from small chromosomes, with not much read information, could be expected to be less trustworthy?

If you have answers for most of my questions in a manual I couldn't find, I want to say: I'm sorry, and could you point me towards it?

Thanks so much for your time, Dr. Davies.

Best,

Pietro