similarity_pyseer generates a matrix with all 0's

[AmayAgrawal]

Hi,

I want to use -lmm to perform GWAS using the linear mixed models. As the -lmm option requires similarity matrix, I saw that there is an option similarity_pyseer which can calculate this similarity matrix. I used this option as follows similarity_pyseer --vcf my_vcf_file.vcf samples.txt > genotype_kinship.tsv to calculate the similarity matrix, but the output matrix generated by this contains all 0's in it and then when I use this matrix to perform GWAS using -lmm, it gives the following error Invalid similarity matrix. Did you use --calc-C?

Can you please help me with this?

[mgalardini]

Could you share your VCF file please?

[AmayAgrawal]

variants.zip

[mgalardini]

Using pyseer 1.3.10 I get a normal looking similarity matrix (see attached):

$ similarity_pyseer --version
similarity 1.3.10

similarity_pyseer --vcf variants.vcf samples.txt > similarity.tsv

similarity.tsv.gz samples.txt.gz

Did you use all samples for your run?

[AmayAgrawal]

I tried the .tsv file that you uploaded to run the GWAS and it works now. I am using the same version of similarity_pyseer (1.3.10)

Yes, I used all samples to construct the similarity matrix. Is it here that I am doing anything wrong?

[mgalardini]

Not that I can see: can you try to run similarity_pyseer --vcf variants.vcf samples.txt > similarity.tsv with the samples.txt file I attached above? Closing the issue anyway, I would say there's no bug that I can see