No input found for *.bedGraph files

[ywang9-npg]

Dear Adam,

I can run the test data smoothly with the script NXF_VER=20.07.1 nextflow run epidiverse/ewas -profile test,conda

But when I run the pipeline use input parameter with script NXF_VER=20.07.1 nextflow run epidiverse/ewas -profile conda --input ./bedGraph --samples samples.tsv, there is an error " No input found for CpG *.bedGraph files: ./bedGraph/CpG" But the bedGraph files are in the folder ./bedGraph/CpG, and I checked that I have the right to read, write and execute these files

Below is the structure of my current work folder . ├── bedGraph │   └── CpG │   ├── sample1_CpG.bedGraph │   ├── sample2_CpG.bedGraph │   ├── sample3_CpG.bedGraph │   ├── sample4_CpG.bedGraph │   ├── sample5_CpG.bedGraph │   └── sample6_CpG.bedGraph └── samples.tsv

bedGraph files I downloaded from https://github.com/epidiverse/datasets/raw/ewas/wgbs/sample1_CpG.bedGraph sample file I downloaded from https://github.com/epidiverse/datasets/raw/ewas/samples.tsv

Thank you! Best regards, Yang

[bio15anu]

Dear Yang,

Sorry for the confusion- the manual is a little misleading. At some point we changed in the directory structure of all the pipelines to integrate better with each other, but I'm not sure how far along we managed to get in that process.

In fact, it looks like EWAS is using the old directory structure, as when I look at the expected input it looks like this: ${params.input}/*/bedGraph/*_CpG.bedGraph

I expect the first * character is meant to capture sample names. So input directory should look something like this:

input_dir
├── sample_1
│   └── bedGraph
│       ├── sample1_CpG.bedGraph
│       ├── sample1_CHG.bedGraph
│       └── sample1_CHH.bedGraph
└── samples.tsv

Sorry again for the confusion!

[ywang9-npg]

thank you very much, it works. Have a nice day！