Could not locate bowtie index (miRDeep2 analysis)

Question

[misslaurenheadley]

Hi,

I have no idea what I've done wrong. I'm trying to run miRDeep2 on some miRNA seq runs. I have linked the bowtie (1) index files to the current directory, and (attempted) to save a path to the originals anyway, I kep getting an error message saying it can't find them.

Below is the script showing what I have in the working directory, the line I tried to execute and everything I got in response.

I then show my path, and that the path I think I used gets me to the files... Any help would be greatly appreciated. Apologies, I'm very new to this so have thrown everything I can think of at it (hence the directory is a pigstye).

  -bash-4.2$ ls
        anaconda3                                                              
        hairpin_hsa_dna.fa  
        miRNA_analysis            
        Homo_sapiens        
        data-shell.zip
        genome.fa 
        genome.1.ebwt
        genome.2.ebwt 
        genome.3.ebwt                                                   
        genome.4.ebwt  
       genome.rev.1.ebwt 
       genome.rev.2.ebwt                  
        Test_1.fq
        dir_mapper_seq_Test_1.fq_3398472581_09_12_2019_t_09_31_42                          
        dir_mapper_seq_Test_1.fq_8799825770_03_12_2019_t_16_37_29 
        index.html                
        GCA_000001405.15_GRCh38_no_alt_analysis_set.fna.bowtie_index.tar  
        mature_hsa_dna.fa   
        shellex

            -bash-4.2$ mapper.pl Test_1.fq -e  -j -k AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC  -l 18 -m -h -p genome  -s reads_collapsed.fa -t reads_collapsed_vs_genome.arf -o 4 -v

            parsing fastq to fasta format
            discarding sequences with non-canonical letters
            clipping 3' adapters
            discarding short reads
            collapsing reads
            mapping reads to genome index
            Could not locate a Bowtie index corresponding to basename "genome"
            Please make sure you used bowtie version 1 to build the index.
            Usual index files have suffix .ebwt

            -bash-4.2$ echo $PATH
            /rds/general/user/lah17/home/anaconda3/bin:/usr/lib64/qt-3.3/bin:/rds/general/user/lah17/home/perl5/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/opt/ibutils/bin:/opt/pbs/bin:/apps/anaconda3/4.5.12/install:/rds/general/user/lah17/home/Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/
            -bash-4.2$ 
            -bash-4.2$ #checking the path to the index...
            -bash-4.2$ cd /rds/general/user/lah17/home/Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/
            -bash-4.2$ ls
            genome.1.ebwt  genome.2.ebwt  genome.3.ebwt  genome.4.ebwt  genome.fa  genome.rev.1.ebwt  genome.rev.2.ebwt

[Barry Digby]

Try

-p genome.

Edit* Checked previous times I have used mapper.pl and nope, that's not it. Maybe provide the path like:

-p ./genome

[misslaurenheadley]

Hi, Ty for the suggestion!

I tried

  -p genome. 

    -bash-4.2$ mapper.pl Test_1.fq -e  -j -k AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC  -l 18 -m -h -p genome. -s reads_collapsed.fa -t reads_collapsed_vs_genome.arf -o 4 -v

    parsing fastq to fasta format
    discarding sequences with non-canonical letters
    clipping 3' adapters
    discarding short reads
    collapsing reads
    mapping reads to genome index
    Could not locate a Bowtie index corresponding to basename genome.
    Please make sure you used bowtie version 1 to build the index.
    Usual index files have suffix .ebwt

Then, out of mild desperation...

  -p "genome.*"

    -bash-4.2$ mapper.pl Test_1.fq -e  -j -k AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC  -l 18 -m -h -p “genome.*” -s reads_collapsed.fa -t reads_collapsed_vs_genome.arf -o 4 -v


    parsing fastq to fasta format
    discarding sequences with non-canonical letters
    clipping 3' adapters
    discarding short reads
    collapsing reads
    mapping reads to genome index
    Could not locate a Bowtie index corresponding to basename "genome.*"
    Please make sure you used bowtie version 1 to build the index.
    Usual index files have suffix .ebwt

Anything else I might try?

Going back to look at PATH..? To see if that might be the issue.

[Barry Digby]

Give it a shot with the full path to the index files, or the relative path (try make a folder for em).

mkdir index
mv *ebwt* index/
cd ../
mapper.pl Test_1.fq -e  -j -k AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC  -l 18 -m -h -p index/genome -s reads_collapsed.fa -t reads_collapsed_vs_genome.arf -o 4 -v

You forgot to try -p ./genome which would be another take on the relative path. If this doesn't work then I'm not sure what to do, except to do the trimming manually (which I have had success with)..

[misslaurenheadley]

Hey,

Thank you for the suggestions. Still hasn't worked. Frustratingly, I have copies that have been trimmed already, I was just trying to do an all in one alignment. Is this what you use for alignment? Back to the drawing board.

bw,

L

[GenoMax]

Did you make the bowtie v.1 indexes yourself or downloaded them from somewhere? Are they known to work?

[misslaurenheadley]

Have figured it out. I missed a full stop when I linked the bowtie sequences. It was looking through the root directory rather than from the current directory I was in.

I learned that black highlight of anything that you ls is not a good thing!

[Barry Digby]

Great! Put up the fixed code, just as a future reference for anyone else coming across the same issue

Answer 1

[misslaurenheadley]

Solution: (I have been using 'Big Data Analysis for Bioinformatics and Biomedical Discoveries, by Shui Qing Ye) There are some errors in the book, but this one was all me,

When linking the the genome and the bowtie index into the current working directory:

  ln -s ./Homo_sapiens/UCSC/hg38/Sequence/WholeGenomeFasta/genome.fa
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.1.ebwt
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.2.ebwt
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.3.ebwt
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.4.ebwt
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.rev.1.ebwt
    ln -s ./Homo_sapiens/UCSC/hg38/Sequence/BowtieIndex/genome.rev.2.ebwt

it may need to be run line by line....

This was based on the files being found from the current directory. then running:

mapper.pl Sample_name.fq -v -q -n -o 4 -u -e -h -m -k AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC -p genome -s reads_ collapsed.fa -t reads_collapsed_vs_genome.arf