search

cumc / xqtl-protocol

Molecular QTL analysis protocol developed by ADSP Functional Genomics Consortium
https://cumc.github.io/xqtl-protocol/
MIT License
39 stars 42 forks source link

Command generators will not run without -n (dryrun option) #262

Open hsun3163 opened 2 years ago

hsun3163 commented 2 years ago

It should be brought to attention that. For our command generator, if run without -n, following error will occurs:

ERROR: [sample_match (sample_match)]: [sample_match]: Output target /mnt/mfs/statgen/snuc_pseudo_bulk/eight_tissue_analysis/MWE/QTL_association/output/data_preprocessing/MWE/sampleSheetAfterQC.filtered.txt does not exist after the completion of step sample_match
[annotation (annotation)]: [annotation]: Output target /mnt/mfs/statgen/snuc_pseudo_bulk/eight_tissue_analysis/MWE/QTL_association/output/data_preprocessing/MWE/phenotype_data/MWE.log2cpm.bed.gz does not exist after the completion of step annotation
[TensorQTL]: Exits with 12 pending steps (TensorQTL, merge_factor_covariate, factor, resid_exp, phenotype_partition_by_chrom, region_list_generation, merge_pca_covariate, projected_sample, pca, related_QC, king, unrelated_QC)

This is expected as the required output was not generated. In contrast, by using the -n option, the sos will automatically generate place holder for those output files, thus escaping the monitoring.

gaow commented 2 years ago

I've got to think about this: we leverage input and output to help with the flow of logic but in this case not the flow of physical files. I am thikning if tehre is a way we can do this such that we take the best of both realms.

hsun3163 commented 2 years ago

I strongly advise against forsaking the input/output statement...