[Bug]: End without warning, but no results are generated？

[AI-10]

System information

Nextflow version 24.10.3
Hardware HCP
Executor local
Container engine Conda
OS Linux
Version of gwas-sumstats-harmoniser v1.1.10

Description of the Issue

I executed the code below. But only the first two steps were completed and no final result was produced.

(base) root@inspur-NF8480M6:/data/gongweiming/software/gwas-sumstats-harmoniser# ./nextflow run EBISPOT/gwas-sumstats-harmoniser --harm -r v1.1.10 --ref /data/gongweiming/software/gwas-sumstats-harmoniser/ref.dbsnp/ --file /data/gongweiming/software/gwas-sumstats-harmoniser/test_data/random_name.tsv -profile standard,conda

Error Message

nextflow.log

Command used and terminal output

No response

First 10 Rows of the Input File

No response

Relevant files

No response

[jiyue1214]

Thank you for using our harmonization pipeline and taking the time to report the issue you encountered. We appreciate your feedback, as it helps us improve the tool.

The issue causing the pipeline termination appears to be related to the ten_percent_counts step. Specifically, the number of chromosomes generated, 22, is less than the default requirement (22 + X, Y, MT = 25), leading to the termination of the pipeline (code).

To resolve this issue, you can try the following:

1. Adjust the chrom parameter in the configuration file located at /root/.nextflow/assets/EBISPOT/gwas-sumstats-harmoniser/config/default_params.config.
2. Use the --chromlist option directly in the command.
3. Ensure your reference includes chromosomes X, Y, and MT. Even if you don't have data for these chromosomes, the pipeline will automatically generate empty files for them.

We acknowledge that the error message in this case is unclear, and we will work on improving it in future updates. Thank you again for your valuable feedback and hope these solutions help.

[AI-10]

Thanks a lot for your reply, this software is really useful before post-GWAS analysis.

I executed the code using the example data in the test_data folder. Same data, but different results.

When I run the following code, it looks normal:
./nextflow run ../gwas-sumstats-harmoniser -profile test,conda

But when I use the below code, an error occurs.
./nextflow run ../gwas-sumstats-harmoniser --harm --ref /data/gongweiming/software/gwas-sumstats-harmoniser/test_data/ --file /data/gongweiming/software/gwas-sumstats-harmoniser/test_data/random_name.tsv -profile standard,conda

Specifying chromosomes or changing default_params.config didn't seem to work

I don't know what happened.

[AI-10]

I would like to know if you have tested the example data following the procedure mentioned in 3.1 General users?

nextflow run EBISPOT/gwas-sumstats-harmoniser \ --ref 'full path to store reference' \ --harm \ --file Full_path_of_the_file_to_be_harmonised or --list path_of_list.txt \ -profile executor,singularity/conda or -profile standard,docker/conda

I'm not sure if it's a problem with the program or the software configuration.

[jiyue1214]

Hi @AI-10,
Thank you for your question! Yes, we thoroughly test the pipeline with both sample and real data before each release.

To confirm, I just re-ran the pipeline on our local HPC after redownloading the entire repository, and it worked successfully. Here is the exact command I used:

nextflow run EBISPOT/gwas-sumstats-harmoniser -r v1.1.10 --harm  --ref $ref/test_data --file $raw_data/random_name.tsv -profile standard,conda --chromlist 1,22

The pipeline completed all steps without issues.

We implemented a check to ensure that the number of chromosomes generated in the second step matches the --chromlist parameter. This is because the workload manager runs all chromosomes in parallel, and we need to confirm that all steps are completed before proceeding to the next stage.

[AI-10]

Hi, jiyue

This may be a software configuration issue.

When I installed nextflow with conda conda install nextflow instead of using the local . /nextflow to perform the analysis, it seemed to work.

I got the fina file, although step 3 is missing. I'm not sure if this result can be used for subsequent analyses.

Many thanks!

[jiyue1214]

Hi @AI-10,

I’m glad to hear it’s finally working!

In Step 2, the pipeline reads the strand orientation of 10% of randomly sampled variants and calculates the strand consistency. If the consistency value is over 0.99, indicating high confidence, the pipeline skips Step 3 (evaluating strand consistency for all variants). This design is intended to save runtime while maintaining accuracy.

Let me know if you have any further questions!

[AI-10]

Hi jiyue,

The programme is now working fine !

Thanks for this great tool !