LinAlgError: Array must not containinfs or NaNs

[yyfyyfzkyzky]

I've been able to run MTAG and get results without any problems before. Only when I use GCST009379 on the GWAS catalog as trait1, it gives me an error no matter which GWAS I use for trait2. The error is as follows:
File "/home/tmj/miniconda3/envs/pgwas/lib/python2.7/site-packages/numpy/linalg/linalg.py", line 218, in _assertFinite
raise LinAlgError("Array must not contain infs or NaNs")
But I went by what others have asked before, and the usual QC. I made sure that the 3 columns pval, z, and n of trait1 did not have NA or Inf, or character formats. But I am still reporting errors.
log.txtGCST009379.ebi-a-GCST006414.log
GWAS summary data: http://diagram-consortium.org/downloads.html (T2D GWAS meta-analysis - Unadjusted for BMI Published in Mahajan et al (2018b))

Please what can I do to fix this problem？

[paturley]

For the trait, can you calculate some summary statistics for the pval, z, and n columns (mean, median, min, max, number of observations)?

…

On Sun, Nov 12, 2023 at 9:36 PM yyfyyfzkyzky ***@***.***> wrote: I've been able to run MTAG and get results without any problems before. Only when I use GCST009379 on the GWAS catalog as trait1, it gives me an error no matter which GWAS I use for trait2. The error is as follows: File "/home/tmj/miniconda3/envs/pgwas/lib/python2.7/site-packages/numpy/linalg/linalg.py", line 218, in _assertFinite raise LinAlgError("Array must not contain infs or NaNs") But I went by what others have asked before, and the usual QC. I made sure that the 3 columns pval, z, and n of trait1 did not have NA or Inf, or character formats. But I am still reporting errors. log.txtGCST009379.ebi-a-GCST006414.log <https://github.com/JonJala/mtag/files/13330183/GCST009379.ebi-a-GCST006414.log> GWAS summary data: http://diagram-consortium.org/downloads.html (T2D GWAS meta-analysis - Unadjusted for BMI Published in Mahajan et al (2018b)) Please what can I do to fix this problem？ — Reply to this email directly, view it on GitHub <#194>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFBUB5MGGG7FTWZZBPTZ5RTYEGBTPAVCNFSM6AAAAAA7IPK5ACVHI2DSMVQWIX3LMV43ASLTON2WKOZRHE4DSNZXGUYDMOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[yyfyyfzkyzky]

chr snpid bpos a1 a2 pval
Min. : 1.000 Length:15085062 Min. : 173 Length:15085062 Length:15085062 Min. :0.0000
1st Qu.: 4.000 Class :character 1st Qu.: 32516319 Class :character Class :character 1st Qu.:0.1900
Median : 8.000 Mode :character Median : 69458674 Mode :character Mode :character Median :0.4500
Mean : 8.679 Mean : 78853802 Mean :0.4597
3rd Qu.:13.000 3rd Qu.:114847980 3rd Qu.:0.7200
Max. :22.000 Max. :249233024 Max. :0.9900
freq n z
Min. :0.0001 Min. :898130 Min. :-38.6667
1st Qu.:0.0014 1st Qu.:898130 1st Qu.: -0.6182
Median :0.0300 Median :898130 Median : 0.1429
Mean :0.2823 Mean :898130 Mean : 0.1295
3rd Qu.:0.5900 3rd Qu.:898130 3rd Qu.: 0.8889
Max. :0.9900 Max. :898130 Max. : 38.6667

min(gcst9379$pval) = 1.136351e-322

[paturley]

What's the length of the n and z columns? (Maybe it's in there somewhere, but the format of this information is very hard to parse.)

…

On Mon, Nov 13, 2023 at 11:06 AM yyfyyfzkyzky ***@***.***> wrote: chr snpid bpos a1 a2 pval Min. : 1.000 Length:15085062 Min. : 173 Length:15085062 Length:15085062 Min. :0.0000 1st Qu.: 4.000 Class :character 1st Qu.: 32516319 Class :character Class :character 1st Qu.:0.1900 Median : 8.000 Mode :character Median : 69458674 Mode :character Mode :character Median :0.4500 Mean : 8.679 Mean : 78853802 Mean :0.4597 3rd Qu.:13.000 3rd Qu.:114847980 3rd Qu.:0.7200 Max. :22.000 Max. :249233024 Max. :0.9900 freq n z Min. :0.0001 Min. :898130 Min. :-38.6667 1st Qu.:0.0014 1st Qu.:898130 1st Qu.: -0.6182 Median :0.0300 Median :898130 Median : 0.1429 Mean :0.2823 Mean :898130 Mean : 0.1295 3rd Qu.:0.5900 3rd Qu.:898130 3rd Qu.: 0.8889 Max. :0.9900 Max. :898130 Max. : 38.6667 min(gcst9379$pval) = 1.136351e-322 — Reply to this email directly, view it on GitHub <#194 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFBUB5KN63FAF65HUNZB56DYEJAPDAVCNFSM6AAAAAA7IPK5ACVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMBYGQ3DENRXG4> . You are receiving this because you commented.Message ID: ***@***.***>

[yyfyyfzkyzky]

length(gcst9379$n)
[1] 15085062
length(gcst9379$z)
[1] 15085062
nrow(gcst9379)
[1] 15085062

[paturley]

Hmm. Does it only have problems when it is trait 1 or also when it is trait 2?

…

On Tue, Nov 14, 2023 at 12:15 AM yyfyyfzkyzky ***@***.***> wrote: length(gcst9379$n) [1] 15085062 length(gcst9379$z) [1] 15085062 nrow(gcst9379) [1] 15085062 — Reply to this email directly, view it on GitHub <#194 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFBUB5NEZ4PAEIWSZPLUUA3YEL467AVCNFSM6AAAAAA7IPK5ACVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMBZGU3DCNBXGM> . You are receiving this because you commented.Message ID: ***@***.***>