FILLINFindHaplotypesPlugin doesn't work

[s.l...@irri.org]

Hello,

I am using TASSEL5.2.8 to do imputation to fill in missing SNPs for GBS data. But there is no output file.

I tried it using hapmap data from chromosome 1. It didn't work. There was no error message as well. But it worked on tutorial data. Any suggestions on that?

Best regards,

Shanshan

The International Rice Research Institute is a member of the CGIAR consortium

[Kelly Swarts]

Hi Shanshan,

Can you send me the system output? If you are using the gui, you can get it by starting the gui from command line (ie start_tassel.pl on unix).

Thanks,

Kelly

[s.l...@irri.org]

Hi Kelly,

Yes, I was using GUI. There was no any output. I selected the target file and donor dir/file basename. But nothing there in output folder. There was no error warning as well. If you let me know your email. I can send you my data file.

[s.l...@irri.org]

Hi,Kelly,

Here is the running log using TASSEL5.0 GUI. 

Chr:1 Alignment Sites:7269 subAlignCnt:1 RealSites:7232 

Chromosome Divisions: 1 start:0 end:7268 

In taxa:374 sites:7269 

SubInAlign Locus:1 StartPos:12978 taxa:374 sites:7269 

Block 0 Inbred and modest coverage:0 

Current Site 0 Current block 0 EndBlock: 113 

Your suggestions are highly appreciated.

Best regards,

Shanshan

On Wednesday, April 15, 2015 at 2:28:02 AM UTC+8, Kelly Swarts wrote:

[Kelly Swarts]

Hi Shanshan,

Could you send me the input parameters you used as well? It should have printed out above the output you forwarded.

Thanks,

Kelly

Thanks,

Kelly

[s.l...@irri.org]

Hi, Kelly,

I used the default parameters in the TASSEL5.0 for the above run I forwarded. The target file was the one i sent to your email just now. As i said, there was no any out put file after the run and no error warning as well.

Your suggestions are highly appreciated.

Thanks and regards,

Shanshan

[Kelly Swarts]

Hi Shanshan,

Your problem is then that the block size is too large. You only have 7269 sites on chromosome one, which is smaller than the default block size of ~8k. The algorithm cannot find taxa that are identical across the whole chromosome to collapse for haplotypes - try changing the default window size to something smaller. Additionally, you shouldn't be filtering before running find haplotypes. All of the sites are informative, and should be included.

Best,

Kelly

[s.l...@irri.org]

Dear Kelly,

Millions thanks for your help.

According to your suggestions, I tried different parameter settings and used the unfiltered hapmap data. If only change the preferred haplotype size to 2000 and min sites to test match to 500 without changing other parameters, it was still not working. Then I set the preferred haplotype size to 1000 and min sites to test match to 200. It worked. One more question, why there is still a lot of missing values after imputation? The proportions of not missing before and after imputation are nearly the same. Any solutions for that.

Many thanks and regards,

Shanshan 

[huana R]

Hi, I am trying to impute my SNP data too. The error sign is FilterGenotypeTable: init: start site:1845 is larger than end site:1844

Is there anyone know how to fix this?

Thank you,

[Kelly Swarts]

Hi Shanshan,

Sorry, i just saw this. There is still missing data because there is missing data in the donor files. FILLINFindHaplotypes collapses genotypes together, but doesn't try to guess genotypes that are missing in all of the taxa collapsed. Part of this is biological - we expect in maize something like 20% of the genome from any given line to not exists in another - but part of it is just that we can't make inference out of nothing.

In your case, if there's very little imputation, I suggest continuing to play with the parameters to try to get your window size down to a level where you expect biologically to share haplotypes between taxa (think about the Ne of your population - the larger it is, the smaller the expected shared haplotype blocks). The other way to increase the probability of finding IBD haplotypes is to increase your sample size - do you have access to other genotyped lines that you can add to your samples for FindHaplotypes?

Best,

Kelly

[Kelly Swarts]

Take a look at your genotype file you are trying to run - the error is saying that you don't have any sites after filtering.

Best,

Kelly