sriramlab / SCORE

SCalable genetic CORrelation Estimator
8 stars 1 forks source link

Running SCORE with bim/fam/bed files per chromosome #3

Open kllrdr opened 2 years ago

kllrdr commented 2 years ago

First, let me congratulate you on SCORE. It is a fantastic tool for the genomics community.

I am running rg analysis with two shared-sample phenotypes. Can I run the analysis with bim, fam, and bed files separate for each chromosome, simply in a for loop? I have tried to run it that way, but I have got the following result (this is specifically for chr21):

Script: ./SCORE/build/SCORE -g chr21 -p phenotypes -c covar -mpheno 1,2 -fill -b 10

phenotype file (both phenotypes are binary): FID IID 1 2 589334 589334 1 0 234578 234578 0 0 1467811 1467811 0 1 456899 456899 1 1

Result: perform GWAS: 0 Binary phenotype: 0 Filling phenotype with mean: 1 Compute heritability: 1 Analysing phenotype 1, 2 Read in 2 phenotypes There are 230883 230883 individuals with no missing phenotypes read in bim file reading functional annotation... Total number of annotation groups: 1 num of random vectors: 100 inverse of exist_ind: 4.3312e-06 4.3312e-06 Read in 13 Covariates.. genotype size: 487409 X 1261158 Total number of blocks: 306initialized Xzb_total.. initialized resid_total.... initialized zb .. 1.17844 -0.365868 -0.0782989 0.446422BLOCK_SIZE: 0 Start reading genotype... Reading chromosome 0 Reading chromosome 1 Reading chromosome 2 Reading chromosome 3 Reading chromosome 4 Reading chromosome 5 Reading chromosome 6 Reading chromosome 7 Reading chromosome 8 Reading chromosome 9 Reading chromosome 10 Reading chromosome 11 Reading chromosome 12 Reading chromosome 13 Reading chromosome 14 Reading chromosome 15 Reading chromosome 16 Reading chromosome 17 Reading chromosome 18 Reading chromosome 19 Reading chromosome 20 Reading chromosome 21 A: -nan -nan -nan 230883 b: -nan -nan V(e): -nan Vp -nan V(G)(0): -nan A: -nan -nan -nan 230883 b: -nan -nan V(e): -nan Vp -nan V(G)(0): -nan total comman samples: 230883 A: -nan -nan -nan 230883 b: -nan -nan rho_g(0): -nan rho_e: -nan gamma_g(0): -nan Start reading genotype... reading chromosome 0 reading chromosome 1 reading chromosome 2 reading chromosome 3 reading chromosome 4 reading chromosome 5 reading chromosome 6 reading chromosome 7 reading chromosome 8 reading chromosome 9 reading chromosome 10 reading chromosome 11 reading chromosome 12 reading chromosome 13 reading chromosome 14 reading chromosome 15 reading chromosome 16 reading chromosome 17 reading chromosome 18 reading chromosome 19 reading chromosome 20 reading chromosome 21 SE(h2g_0(0)): nan SE(h2g1(0)): nan SE(rg(0,1)(0)): nan

Thank you so much in advance!

ariel-wu commented 2 years ago

Hi! Thanks for reaching out! We updated the software with an output option, "-o [output name]", where you can easily with output redirected. "nan" indicates a numerical issue, which could happen for many reasons:

  1. Phenotypes have undefined values. The missing values should be marked as "NA".
  2. There are too many missing values in the phenotypes, thus using the option "-fill" will end up with an issue. Try not to use this option.
  3. There are variants that are too rare or too much missingness. Please try providing quality control over genotype.