mtokuyama / ervmap Goto Github PK

Here's my command, and ${i} is the fastq file

erv_genome.pl -stage 1 -stage2 6 -fastq /${i}.fastq.gz

And the software prints

Usage:
    erv_se_genome_v2.pl [options]

       Options:
        -h, -? --help            brief help message
        -t, --test               print out commands without run
        --btrim                  path to btrim (http://graphics.med.yale.edu/trim/)
        --tophat                 path to tophat
        --bwa                    path to bwa
        --samtools               path to samtools
        --filter                 path to filter (parse_bam.pl)
        --bedtools               path to bedtools

        --genome                 path to bwa human genome index
        --genome_Bowtie2         path to Bowtie2 human genome index
        --bed                    path to bed file of ERVs
        --genomefile             path to genome size file
        --gtf                    path to gtf file of human gene annotation
        --transcriptome          path to known transcriptome, used by tophat2
        --adaptor                path to adaptor files, used by btrim (http://graphics.med.yale.edu/trim/)

        --fastq                  fastq file
        --stage                  start stage (see below)
        --stage2                 end stage (see below)


        Stages:
          1                      trim adaptors and low quality regions
          2                      map reads using bwa
          3                      sort bam file
          4                      count reads mapped to ERVs
          5                      map reads using  tophat2
          6                      counts reads mapped to genes

Any idea on this?

I am trying this tool, part till alignment ran fine but at htseq count step found the following error
/usr/bin/python2.7 htseq.script.count not found.

Hello,

Do you have a recommended code for differential expression of ERVS? Here are a few issues I am not sure about:
-if ERVS are normalised with your script, and I use the normalised ERV matrix in DESeq2, it is worth noting that DEseq function by default will estimate size factors. This will effectively result in double normalisation. In order to overcome this, one could not use DEseq, but its constituents:
#dds <- estimateSizeFactors(dds)
dds <- estimateDispersions(dds)
dds <- nbinomWaldTest(dds)
By skipping estimateSizeFactors, the double normalisation will be avoided.

Another option is to NOT use your script provided for normalisation, and then to load the sizeFactors obtained by cellular counts. If sizeFactors exist, estimateSizeFactors will use the existing ones and not attempt to recalculate.

Yet another option would be to merge the raw cellular and erv matrices, and go straight to DEseq where they DEseq function will do estimateSizeFactors on the entire matrix of merger cellular and era counts.

Would you be able to comment on these options?
Thank you,
Regards

Hi @mtokuyama I didn't detect any ERVmap.
I used docker https://hub.docker.com/r/eipm/ervmap

 singularity exec \
    --bind ${dir}/raw:/data:ro \
    --bind ${index}:/genome:ro \
    --bind ${erv_bed}:/resources:ro \
    --bind ${dir}/ERV.map/results/${i}:/results \
    /home/zhou/raid/singularity/ervmap.1.2.1.sif /scripts/ERVmapping.sh \
    --read1 ${dir}/raw/${i}_1.fastq.gz --read2 ${dir}/raw/${i}_2.fastq.gz \
    --output ${i} -m ALL \
    --cpus 12 --limit-ram 151290751290 -d

But those two samples didn't detect any ERV.

SRR3083781Log.final.out.txt
SRR3083781ERVresults.txt
SRR10899987Log.final.out.txt
SRR10899987ERVresults.txt
Could you please help me to check this issue?
Best,
JG

Hi,

I run the bioinformatics core at La Jolla Institute for Immunology (lji.org), and one of our researchers is interested in running your tool. I've had a quick look through the repo and was unable to find any documentation in terms of installing or using the code. Please let me know if you can provide this.

Thanks very much,

Jason

Hi Maria,

Thank you very much for developing this robust tool!

I am conducting the differential ERVs analysis. I have a question regarding if we should take gene expression into consideration.
Which way do you think is better?

1. calculate the differential ERVs based on the count of ERVs only.
2. calculate the differential ERVs based on the input file of the count of ERVs and genes (that basically means i will cat the two count files and conduct the differential ERV analysis with DEseq2)

Thank you so much!

Best,

Zunpeng

Hi!
I think ERVmap.bed don't provide the ERV family information. Although I find some IDs contain description of HERV-K and HERV-W, most (3031 in 3220) IDs are number ID. Do you have any suggestions about where I can find family information of these HERVs? Such as database. Or do you have a file containing family information that you would like to share?
Thank you!
Zelin

How was the ERVmap.bed generated? I would like to cluster the ERVs into groups in a downstream analysis of the ERVs but am not sure where the names in the bed file came from. Thanks!

Error: Sorted input specified, but the file ERVmap2.bed has the following out of order record
chr1    12704128        12712798        5818    500     +

Hi,

Is it convenient to share the --genomefile of hg38, I use the genome size file generated by the STAR soft, while the ERVmap shows the chromose does not match, so I add "chr" to each line of the file, while then ERVmap shows that some line out of order record. Any suggestions? Thanks a lot!

Best,

Hello,

The requested version of Btrim is not available any more - not via Conda and not in the link you are providing on GitHub (http://graphics.med.yale.edu/trim/). I tried with other versions, but the parameters ERVmap is sending to Btrim are not compatible withe available versions of Btrim.

Any idea how to resolve it?
Thank you

Hi,

May I know how to filter the results, as some item has few overlaps(eg. < 100), is there a threshold(eg. >20 bp) for the result to be considered as a truly detected ERV.

Thanks so much for your guidance!