Closed Estersf closed 3 years ago
now2014
commented
3 years ago Hi Estersf,
Thanks for your bug report! This error is due to the mismatch of filenames in the LD reference panel; I've fixed it. You may need to download the updated code and re-install it.
As for the warning message, are you using two different GWAS summarise? The missing rate is related to the proportion of overlapped SNPs between GWAS summary statistics and the reference panel.
Best regards, now2014
Estersf
commented
3 years ago Hi, Thank you for your quick response. I have downloaded the updated code and I get this error on re-installation:
No, I am using the same GWAS summarise.
Thank you Ester
now2014
commented
3 years ago Sorry, I left a typo in the code. You can retry it now.
Estersf
commented
3 years ago Now it works fine. Thank you very much. However, I have another question that I would like to share with you, in case you could help me with any suggestions or advice, please.
Although I don’t get any warning message in the “HDL.data.wrangling.R” step:
Program starts on Mon Jul 5 13:06:57 2021
Loading GWAS summary statistics from /home/ester/Documents/Ester/SCZ/METAL/STDERR_OFF/META_chrs_bp_N_rsID_good.txt
Data are loaded successfully. Data wrangling starts.
Data wrangling completed.
763617 out of 769306 (99.26%) SNPs in reference panel are available in GWAS.
The output is saved to /media/ester/Elements/SCZ_AAO/HDL/Outputs/test.hdl.rds
The log is saved to /media/ester/Elements/SCZ_AAO/HDL/Outputs/test.txt
In the “HDL.run.R” step, the proportion of overlapped SNPs between GWAS summary statistics and the reference panel has changed:
Function arguments: gwas.df=/media/ester/Elements/SCZ_AAO/HDL/Outputs/test.hdl.rds LD.path=/media/ester/Elements/SCZ_AAO/HDL/UKB_imputed_hapmap2_SVD_eigen99_extraction output.file=/media/ester/Elements/SCZ_AAO/HDL/Outputs/gwas_UKB_imputed_hapmap2.Rout
HDL: High-definition likelihood inference of genetic correlations and heritabilities (HDL) Version 1.4.0 (2021-04-15) installed Author: Zheng Ning, Xia Shen Maintainer: Zheng Ning zheng.ning@ki.se Tutorial: https://github.com/zhenin/HDL Use citation("HDL") to know how to cite this work.
Analysis starts on Mon Jul 5 13:07:23 2021
749497 out of 769306 (97.43%) SNPs in reference panel are available in the GWAS.
Warning: More than 1% SNPs in reference panel are missed in the GWAS. This may generate bias in estimation. Please make sure that you are using correct reference panel.
Integrating piecewise results
Continuing computing standard error with jackknife
Heritability of phenotype 1: 0.00e+00 (0.00e+00)
P: NA
Warning: Heritability of the trait was estimated to be 0, which may due to:
1) The true heritability is very low;
2) The sample size of the GWAS is too small;
3) Many SNPs in the chosen reference panel are absent in the GWAS;
4) There is a severe mismatch between the GWAS population and the population for computing reference panel
Analysis finished at Mon Jul 5 13:11:08 2021
The results were saved to /media/ester/Elements/SCZ_AAO/HDL/Outputs/gwas_UKB_imputed_hapmap2.Rout
I don’t see the point of this loss of overlapped SNPs. Do you have any suggestions on this issue? Thank you! Ester
Estersf
commented
3 years ago Also, as a complementary doubt/curiosity, I noticed that if for the same GWAS, I use different panels in the "HDL.data.wrangling.R" step and in the "HDL.run.R" step, the resulting heritability is higher (h2=0.21) than using the same panel in both steps (h2=0). However, the percentage of overlapping SNPs is much lower (72.78%) than using the same panel in both steps (97.43%).
Do you have an explanation for this increase in heritability? Any ideas/comments would be appreciated.
Thank you for your time
Ester
now2014
commented
3 years ago In the HDL.run.R step, a parameter called fill.missing.N with default value = NULL, which means HDL will filter out variants whose Z-scores are NAs. But these variants are not filtered in the HDL.data.wrangling.R step. So you got different percentages of overlapped variants with the same LD reference panel.
You shouldn't use different reference panels in HDL.data.wrangling.R and HDL.run.R steps; because you'll filter the variants TWICE. The HDL.data.wrangling.R not only formats the GWAS result into designated columns but also removes variants absent in the LD reference panel, which is also performed in the HDL.run.R step.
When the missing rate is high, you may try a lower eigen.cut. Meanwhile, HDL is not recommended for heritability estimation.
(#14)
Hi,
I am trying to calculate the heritability from my own summary statistics data. Using the UKB_imputed_SVD_eigen99_extraction panel I get a warning message “Warning: More than 1% SNPs in reference panel are missed in the GWAS. This may generate bias in estimation. Please make sure that you are using correct reference panel”, so I switched to the UKB_imputed_hapmap2_SVD_eigen99_extraction panel, which does not give me any warning message in the “HDL.data.wrangling.R” step:
Rscript /home/ester/Downloads/HDL/HDL.data.wrangling.R \
Attaching package: ‘dplyr’
The following objects are masked from ‘package:stats’:
The following objects are masked from ‘package:base’:
Attaching package: ‘data.table’
The following objects are masked from ‘package:dplyr’:
Program starts on Thu Jul 1 16:06:29 2021 Loading GWAS summary statistics from /home/ester/Documents/Ester/SCZ/METAL/STDERR_OFF/META_chrs_bp_N_rsIDgood.txt Data are loaded successfully. Data wrangling starts. Warning message: `rename()
was deprecated in dplyr 0.7.0. Please userename()instead. This warning is displayed once every 8 hours. Calllifecycle::last_warnings()` to see where this warning was generated. Data wrangling completed. 763617 out of 769306 (99.26%) SNPs in reference panel are available in GWAS.The output is saved to /media/ester/Elements/SCZ_AAO/HDL/Outputs/test.hdl.rds The log is saved to /media/ester/Elements/SCZ_AAO/HDL/Outputs/test.txt
However, I get an error in the “HDL.run.R” step, which I am not able to solve:
Rscript /home/ester/Downloads/HDL/HDL.run.R \
Loading GWAS ...
HDL: High-definition likelihood inference of genetic correlations and heritabilities (HDL) Version 1.4.0 (2021-04-15) installed Author: Zheng Ning, Xia Shen Maintainer: Zheng Ning zheng.ning@ki.se
Tutorial: https://github.com/zhenin/HDL
Use citation("HDL") to know how to cite this work.
Analysis starts on Thu Jul 1 16:06:48 2021 749497 out of 769306 (97.43%) SNPs in reference panel are available in the GWAS.
Warning: More than 1% SNPs in reference panel are missed in the GWAS. This may generate bias in estimation. Please make sure that you are using correct reference panel.
Error in readChar(con, 5L, useBytes = TRUE) : cannot open the connection Calls: HDL.h2 -> load -> readChar In addition: Warning message: In readChar(con, 5L, useBytes = TRUE) : cannot open file '/media/ester/Elements/SCZ_AAO/HDL/UKB_imputed_hapmap2_SVD_eigen99_extraction/': it is a directory Execution halted
Could you suggest me some advice on how to solve this error, please? I am new to use this tool and I also do not understand why, in the second step, more than 1% SNPs in reference panel are missed in the GWAS, as I am using the same panel in both steps.
I look forward to your reply and thank you for this tool.
Best regards,
Ester