SNP with low variance

[lanjiangVUMC]

I am new to Regenie and trying to test it on my dataset. here's my setting in step1: regenie \ --step 1 \ --bed chr22_prune \ --keep lowtg_samples.txt \ --covarFile pheno.txt \ --phenoFile pheno.txt \ --phenoCol lowtg \ --covarCol age,PC1,PC2 \ --bsize 100 \ --bt --lowmem \ --lowmem-prefix tmp_rg \ --out fit_bin_out

And the log file is Log of output saved in file : fit_bin_out.log

Options in effect: --step 1 \ --bed chr22_prune \ --keep lowtg_samples.txt \ --covarFile pheno.txt \ --phenoFile pheno.txt \ --phenoCol lowtg \ --covarCol age,PC1,PC2 \ --bsize 100 \ --bt \ --lowmem \ --lowmem-prefix tmp_rg \ --out fit_bin_out

Fitting null model

• bim : [chr22_prune.bim] n_snps = 14049
• fam : [chr22_prune.fam] n_samples = 35024 -keeping only individuals specified by --keep -number of genotyped individuals remaining in the analysis = 6258
• bed : [chr22_prune.bed]
• phenotypes : [pheno.txt] n_pheno = 1 -dropping observations with missing values at any of the phenotypes -number of phenotyped individuals with no missing data = 6258
• covariates : [pheno.txt] n_cov = 3 -number of individuals with covariate data = 6258
• number of individuals used in analysis = 6258
• case-control counts for each trait:
  • 'lowtg': 486 cases and 5772 controls -fitting null logistic regression on binary phenotypes...done (3ms) -residualizing and scaling phenotypes...done (0ms)

• threads : [11]

• block size : [100]

• blocks : [141] for 14049 variants

• CV folds : [5]

• ridge data_l0 : [5 : 0.01 0.25 0.5 0.75 0.99 ]
• ridge data_l1 : [5 : 0.01 0.25 0.5 0.75 0.99 ]
• approximate memory usage : 37MB
• writing level 0 predictions to disk -temporary files will have prefix [tmp_rg_l0_Y] -approximate disk space needed : 34MB
• setting memory...done

Chromosome 22 block [1] : 100 snps (3ms) -residualizing and scaling genotypes...done (7ms) -calc working matrices...done (2ms) -calc level 0 ridge...done (10ms) block [2] : 100 snps (3ms) -residualizing and scaling genotypes...done (6ms) -calc working matrices...done (2ms) -calc level 0 ridge...done (10ms) block [3] : 100 snps (3ms) -residualizing and scaling genotypes...done (5ms) -calc working matrices...done (2ms) -calc level 0 ridge...done (10ms) block [4] : 100 snps (2ms) -residualizing and scaling genotypes...ERROR: !! Uh-oh, SNP 22:10732714:C:T has low variance (=0.000000).

I calculated the allele frequencies for the subset of 6258 individuals and the minimal maf is 0.03. I am not sure why the program is giving error. What would be your suggestions? Thank you very much!

[Ojami]

If you're sure that min AAF in the subset of individuals analyzed by REGENIE is 0.03, please check if the problematic variant is highly correlated with your covariates (probably PCs). Test if this error happens if you remove PC1,PC2 from --covarCol.

[lanjiangVUMC]

I re-ran it without covariates and it still gave me the same type of error. The maf of the variant is 0.5. Will this be a problem? Thanks. 22:10732714:C:T C T 0.5 12516