WGBS: error when using custom genome: Not a directory

[rmblazquez]

Greetings, I'm trying to run the EpiDiverse WGBS pipeline using a custom reference plant genome, but the pipeline seems to have trouble recognizing the reference genome directory files and subfolders.

The command I'm running is:

NXF_VER=20.07.1 nextflow run epidiverse/wgbs -profile conda --input $readsDir --reference $genome --trim --fastqc

where $readsDir is the path to Illumina PE reads from a WGBS (a directory), and $genome is, as described in the help, the path to the FASTA file. I run the command within a conda environment with all epidiverse dependencies accomodated.

I followed the custom reference section to prepare all the files and subfolders needed for the reference genome in the ecic_genome directory. This is what ls * within the genome directory looks like:

~/epidiv/ecic_genome/ecic_genome.fa  ecic_genome/ecic_genome.fa.fai

~/epidiv/ecic_genome/index:
ecic_genome.ebm

~/epidiv/ecic_genome/lambda:
lambda.ebm

The genome sequence with the .fa extension and its index generated with samtools faidx are in the main directory, while the erne-bs5 index is in the index folder, and the reference genome plus lambda phage index genome is in the lambda folder. Since I'm using only erne-bs5 for alignment, I'm not including .ctidx or .gaidx indexes.

However, when I set the variable $genome from the command line to ~/epidiv/ecic_genome/ecic_genome.fa, I get the following error:

Nextflow 24.04.3 is available - Please consider updating your version to it
WARNING: An illegal reflective access operation has occurred
WARNING: Illegal reflective access by org.codehaus.groovy.reflection.CachedClass (file:~/anaconda3/envs/epidiv/share/nextflow/capsule/deps/org/codehaus/groovy/groovy/2.5.11/groovy-2.5.11.jar) to method java.lang.reflect.AnnotatedElement.lambda$getDeclaredAnnotationsByType$0(java.lang.annotation.Annotation,java.lang.annotation.Annotation)
WARNING: Please consider reporting this to the maintainers of org.codehaus.groovy.reflection.CachedClass
WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
WARNING: All illegal access operations will be denied in a future release
N E X T F L O W  ~  version 20.07.1
Launching `epidiverse/wgbs` [determined_tesla] - revision: 17d09d16d8 [master]
~/epidiv/ecic_genome/ecic_genome.fa/lambda: Not a directory

 -- Check script '~/.nextflow/assets/epidiverse/wgbs/main.nf' at line: 203 or see '.nextflow.log' file for more details

I follow as the error suggests and I check the .nextflow.log file, which shows me again the "Not a directory" error:

Jul-24 11:44:21.422 [main] DEBUG nextflow.script.ScriptRunner - > Launching execution
Jul-24 11:44:21.468 [main] DEBUG nextflow.Session - Session aborted -- Cause: ~/epidiv/ecic_genome/ecic_genome.fa/lambda: Not a directory
Jul-24 11:44:21.479 [main] ERROR nextflow.cli.Launcher - ~/epidiv/ecic_genome/ecic_genome.fa/lambda: Not a directory
java.nio.file.FileSystemException: ~/epidiv/ecic_genome/ecic_genome.fa/lambda: Not a directory

When I check ~/.nextflow/assets/epidiverse/wgbs/main.nf, I go to line 163, // VALIDATE INITIAL PARAMETERS, to check how the parameter validation is done:

    // VALIDATE INITIAL PARAMETERS
    ParameterChecks.checkParamsEpi(params)
    include { check_ref_errors } from './lib/functions.nf' params(reference: params.reference, thlaspi: params.thlaspi, populus: params.populus, fragaria: params.fragaria, nolambda: params.noLambda)
    (fasta, fai, ebm_path, ctidx_path, gaidx_path) = check_ref_errors(params.reference, params.thlaspi, params.fragaria, params.populus, params.noLambda)
}

else {

    // VALIDATE INITIAL PARAMETERS
    ParameterChecks.checkParamsLocal(params)
    fasta = file("${params.reference}", checkIfExists: true, glob: false)
    fai = file("${params.reference}.fai", checkIfExists: true, glob: false)
    ebm_path = params.noLambda && params.split == "${baseDir}/data/lambda.fa" ? "${params.reference}/index/*.ebm" : "${params.reference}/lambda/*.ebm"
    ctidx_path = params.noLambda && params.split == "${baseDir}/data/lambda.fa" ? "${params.reference}/index/*.ctidx" : "${params.reference}/lambda/*.ctidx"
    gaidx_path = params.noLambda && params.split == "${baseDir}/data/lambda.fa" ? "${params.reference}/index/*.gaidx" : "${params.reference}/lambda/*.gaidx"
}

Could it be that variable params.reference is sometimes implied to be a file and sometimes a directory in this code? Do I need also the .ctidx and .gaidx files to run the analysis even if I only use erne-bs5? Is there any way to manually change the params.referece variable to suit the custom reference genome file structure?

I'm not running the command within the genome folder, but in a different directory. I'm also defining my paths with the complete paths (I'm using ~ in this post instead of my home path). I also can verify that the epidiverse/wgbs test run was succesfull with the sample dataset.

Let me know if there is anything else you need to track this error, or if there are some mistakes on my behalf.

Best regards,

Rubén MB