Nextflow: [E::bwa_idx_load_from_disk] fail to locate the index files

Question

[Eliveri]

I have a nextflow workflow which I get the following error:
(I tried the path /path/genomes/human.fa and the path to bwa index files exists.)

Command error:
  [E::bwa_idx_load_from_disk] fail to locate the index files

The nextflow.config looks like this: I think the error may be due to the way the path variable is written in the nextflow.config or main.nf?

params {

refdir = "./genomes"
ref = "${params.refdir}/human.fa"

And main.nf looks like this for the bwa process:

ref = file(params.ref)
refdir = file(params.refdir)

... 
process bwa_align {

    tag "align ${pair_id}"

    publishDir "${params.outdir}/$pair_id"

    input:
    tuple val(pair_id), path(paired_reads), path(unpaired_reads)

    output:
    tuple val(pair_id), path("${pair_id}.sam")

    script:

    """
    bwa mem -t ${params.max_threads} -M $ref ${paired_reads} > ${pair_id}.sam
    """
}

[dthorbur]

It looks like the bwa index files are not being loaded into the environment. If you are in a container, are you sure the bind path is appropriate? You may need to use export SINGULARITY_BIND="/path/to/project/directory" in the beforeScript nextflow argument. Additionally, maybe it's unnecessary to add value channels to your input declaration, but I always put all necessary files and paths there including the index files. It makes checking logically checking for errors easier. So, if I needed the ref.dict and ref.fasta.fai index files, I would do something like this:

ref_genome = file( params.RefGen, checkIfExists: true )
ref_dir    = ref_genome.getParent()
ref_name   = ref_genome.getBaseName()
ref_dict   = file( "${ref_dir}/${ref_name}.dict", checkIfExists: true )
ref_index  = file( "${ref_dir}/${ref_name}.{fasta|fna|fa}.fai", checkIfExists: true )

Channel
  .fromFilePairs("${params.ProcBamDir}/*{bam,bai}") { file -> file.name.replaceAll(/.bam|.bai$/,'') }
  .ifEmpty { error "No bams found in ${params.ProcBamDir}" }
  .map { ID, files -> tuple(ID, files[0], files[1]) }
  .set { processed_bams }

process Some_Process {
  input:
  set SampleID, path(bam), path(bai) from processed_bams
  path ref_genome
  path ref_index
  path ref_dict

  [rest of process...]
}

[Maxime Garcia]

what do you specify in your apptainer profile? I'm guessing you might be missing for it. This is an example that is used in the nf-core/configs for a specific HPC with singularity: https://github.com/nf-core/configs/blob/550f4745b61449cd2a57ac7ee5f232a87dd6450a/conf/abims.config#L10-L11 But it might give you some ideas

[Eliveri]

Hi the apptainer in nextflow.config looks like this

apptainer {
    conda.enabled           = false
    apptainer.enabled       = true
    apptainer.autoMounts    = true
    docker.enabled          = false
    process.container       = 'file://image.sif'
}

Answer 1

[Pierre Lindenbaum]

refdir = "./genomes"

use a full path for refdir

[Eliveri]

I ended up passing in refdir as an input into my processes. Making sure full paths are used resolved most of the issues I encountered! Thank you.