Getting nan values for bC bC_se bC_pval in mtCOJO results

[Mahantesh-Biradar]

I'm running mtCOJO using following script and getting nan values for bC , bC_se, bC_pval in the mtCOJO results.

The run has successfully completed but much helpful info in the log.

Can you please suggest what might be going wrong?

Thanks in advance

[anglixue]

Hi, Can you attach the full log file indicating the version number and command you used? and also the first few lines of your two GWAS summary statistics?

[Mahantesh-Biradar]

Hi, sorry for the delay. Here you go:

1. Command --bfile Imp_chr1 --mtcojo-file /mtcojo_summary_data.list --ref-ld-chr /Reference/eur_w_ld_chr/ --w-ld-chr /Reference/eur_w_ld_chr/ --out /AC_conditioned_on_BA --threads 7

2. version v1.94.1 Linux

3. First few lines of GWAS summary stats: SNP A1 A2 freq b se p N rs10875231 t g 0.2424 0.024 0.011 0.02967 233017 rs186077422 a g 0.0055 -0.4012 0.2296 0.08058 233017 rs6678176 t c 0.2989 0.021 0.0104 0.04241 233017 rs78286437 t c 0.9437 -0 0.0219 0.9995 233017 rs144406489 a g 0.9797 0.0205 0.0359 0.5672 233017

SNP A1 A2 freq b se p N rs10875231 T G 0.2465 -0.4707 0.26841694 0.079120003 44039 rs6678176 T C 0.3052 -0.54269999 0.25082907 0.030230001 44039 rs78286437 C T 0.0586 -0.9404 0.53560597 0.078749999 44039 rs78642210 T C 0.0596 -0.96079999 0.54203492 0.075910002 44039 rs113470118 G A 0.0596 -0.83170003 0.51144195 0.1037 44039

[anglixue]

Hi, Can you show the full log file in which I can see all the intermediate messages for troubleshooting? Thanks! I suspect it's either a format issue or the LDSC estimate is problematic.

[Mahantesh-Biradar]

Hi, I'm running it chromosomewise and there is no errors in the log files. Here is log from one of the chr files:

mtCOJO analysis ... There are 277109 SNPs in common between the target trait and all the covariate trait(s). Saving the mtCOJO analysis results of 277109 remaining SNPs to AC_conditioned_on_BA1.mtcojo.cma] ... mtCOJO analysis completed.

[anglixue]

Hi, is this the full log file your obtained?

A full log file of mtCOJO analysis should include the following lines:

Reading the PLINK FAM files Reading GWAS summary data from Reading PLINK BED files Univariate LD score regression analysis to estimate SNP-based heritability ... Bivariate LD score regression analysis to estimate genetic correlation between each pair of traits ... GSMR analysis for covariate #1 (XXX) ... GSMR analysis for covariate #2 (XXX) ... mtCOJO analysis ...

Do you have those information in your log file?

[Mahantesh-Biradar]

Hi, I do have these in my log as follows:

Reading the PLINK FAM files Reading GWAS summary data from Univariate LD score regression analysis to estimate SNP-based heritability ... GSMR analysis for covariate #1 ... 4 index SNPs are obtained from the clumping analysis with p < 5.0e-08 and LD r2 < 0.05. Warning: Not enough SNPs to perform the GSMR analysis. At least 10 SNPs are required. bxy is estimated from rg. bxy nan nan bxy nan nan GSMR analysis for covariate #1 completed.

mtCOJO analysis ... There are 277109 SNPs in common between the target trait and all the covariate trait(s). Saving the mtCOJO analysis results of 277109 remaining SNPs to ....

[anglixue]

Hi, As the log file said, there is not enough SNPs to perform GSMR analysis. I would suggest you to lower the LD r2 threshold to 0.1 to see if you can clump 5 SNPs. I usually won't recommend lowering the GWAS p-val threshold, but if you can't find any larger GWAS, try lowering it to 5E-7 or 1E-5.

[Mahantesh-Biradar]

Hi Angli,

Thank you very much for your suggestion. Let me try this and get back to you.