/ikra_dev

RNAseq pipeline centered on Salmon

Primary LanguageJupyter NotebookMIT LicenseMIT

DOI

ikra v1.2.1 -RNAseq pipeline centered on Salmon-

A gene expression table (gene × sample) is automatically created from the experiment matrix. The output can be used as an input of idep. Ikra is an RNAseq pipeline centered on salmon.

日本語ドキュメントはこちら

Usage

Usage: ikra.sh experiment_table.csv species \
        [--test, --fastq, --help, --without-docker, --udocker --protein-coding] \
        [--threads [VALUE]][--output [VALUE]]\
        [--suffix_PE_1 [VALUE]][--suffix_PE_2 [VALUE]]
  args
    1.experiment matrix(csv)
    2.reference(human or mouse)

Options:
  --test  test mode(MAX_SPOT_ID=100000).(dafault : False)
  --fastq use fastq files instead of SRRid. The extension must be foo.fastq.gz (default : False)
  -u, --udocker
  -w, --without-docker
  -pc, --protein-coding use protein coding transcripts instead of comprehensive transcripts.
  -t, --threads
  -o, --output  output file. (default : output.tsv)
  -s1, --suffix_PE_1    suffix for PE fastq files. (default : _1.fastq.gz)
  -s2, --suffix_PE_2    suffix for PE fastq files. (default : _2.fastq.gz)
  -h, --help    Show usage.
  -v, --version Show version.
  • test option limits the number of reads to 100,000 in each sample.
  • udocker mode is for server environments that can only use User privileges. For more information https://github.com/indigo-dc/udocker.
  • without-docker mode works with all tools installed. Not recommended.
  • protein-coding mode restricts genes to protein coding genes only.
  • threads
  • output is output.tsv by default.
    experiment matrix should be separated by commas (csv format).

SRR mode

name SRR Layout condition1 ...
Treg_LN_1 SRR5385247 SE Treg ...
Treg_LN_2 SRR5385248 SE Treg ...

fastq mode

name fastq(PREFIX) Layout condition1 ...
Treg_LN_1 hoge/SRR5385247 SE Treg ...
Treg_LN_2 hoge/SRR5385248 SE Treg ...
  • Denote names by connecting conditions and replicates with underscores. See idep's Naming convention in detail.
  • The first three columns are required.
  • If you want to use your own fastq file, add --fastq option. Ikra supports only .fq, .fq.gz, .fastq and fastq.gz.
  • fastq file specifies path excluding fastq.gz or _1.fastq.gz and _2.fastq.gz. For example, hoge/SRR5385247.fastq.gz is described as hoge/SRR5385247.
  • If suffix is not _1.fastq.gz or _2.fastq.gz, add -s1 and -s2 options.
  • It is impossible for docker to specify a hierarchy above the execution directory, such as ../fq/**.fastq.gz, but it can be avoided by pasting a symbolic link. bonohu blog

Output

  • output.tsv(scaledTPM)
  • multiqc_report.html : including fastQC reports and mapping rate of salmon(mapping rate for transcripts)

output sample

Treg_LN_1 Treg_LN_2
0610005C13Rik 0 0
0610006L08Rik 0 1
0610009B22Rik 4 10
...

Specification

Bug 2019/04/30

A serious bug was reported in the tximport_R.R and fixed. In the older version, Salmon's output and multiqc reports were correct and sometimes output.tsv were disturbed. Please update Ikra to the latest version. If you are using the old version(<1.1.1), please update and re-run ikra. We apologize for the inconvenience.

Install

All you need is git clone ikra, and install docker or udocker(v1.1.3). No need for installing plenty of softwares! If you don’t want to use docker (or udocker), you must install all softwares by yourself and use —-without-docker option.

$ git clone https://github.com/yyoshiaki/ikra.git

test

Illumina trim_galore ver.

SE

SRR mode

$ cd test/Illumina_SE && bash ../../ikra.sh Illumina_SE_SRR.csv mouse --test -t 10

fastq mode

You can execute it after you execute SRR mode. (That is because you don’t have fastq files.)

$ cd test/Illumina_SE && bash ../../ikra.sh Illumina_SE_fastq.csv mouse --fastq -t 10

PE

SRR mode

$ cd test/Illumina_PE && bash ../../ikra.sh Illumina_PE_SRR.csv mouse --test -t 10

fastq mode

You can execute it after you execute SRR mode. (That is because you don’t have fastq files.)

$ cd test/Illumina_PE && bash ../../ikra.sh Illumina_PE_fastq.csv mouse --fastq -t 10

For Mac Users

Dr.Ota(DBCLS) solved the problem that salmon doesn’t work on Mac. The cause of the problem is that Docker is allocated only 2GB by default on Mac. The problem will be solved by allocating sufficient memory space(>=8Gb) for Docker, and applying and restarting Docker.

img img

ikra pipeline

Tips

You can find SRR data so quickly in http://sra.dbcls.jp/

Issue

Please refer to issue

Releases

Please refer to Relases

  • add support for udocker
  • add setting of species
  • gtf and transcript file from GENCODE
  • salmon
  • trimmomatic(legacy)
  • trim_galore!
  • tximport
  • fastxtools(for Ion)
  • judging fastq or SRR(manual)
  • introduce "salmon gcbias correction"
  • salomn validateMappings
  • pigz(multithread version of gzip)
  • fasterq-dump
  • cwl development is in progress
  • rename to "ikra"
  • protein coding option

Legacy

Moved the flow using trimmomatic to ./legacy

Reference

Development of cwl ver.

2019/03/22 https://youtu.be/weJrq5QNt1M We tried developing it because Mr.Michael visited Japan. For now, cwlnized trim_galore and salmon in PE.

cd test/cwl_PE && bash test.sh

sorce and reference ー cwl_tools

Citation

Hiraoka, Y., Yamada, K., Kawasaki, Y., Hirose, H., Matsumoto, K., Ishikawa, K., & Yasumizu, Y. (2019). ikra : RNAseq pipeline centered on Salmon. https://doi.org/10.5281/ZENODO.3352573