Exploring a large cancer cell line RNA-sequencing dataset with k-mers

Scripts and data files for reproducing the publication results

1. Index preparation

The CCLE Reindeer index

The core of the study is the CCLE index built by Reindeer, comprising 1019 samples. It is too large to be stored here (236 GB). A better way is to ask us to transfer it to you. But if you want build your own CCLE index, here is the procedure.

NB: this procedure is suited to Reindeer V 1.02 or earlier.

• download all 1019 fastq files (11 TB), the sra-CCLE_metadata.tsv file contain the list and links to download the selection of fastq files.
• trim the fastq with cutadatp -q 10,10 -m 31 -p out2.fastq in1.fastq in2.fastq
• build the index (prefer --disk-mode) using trimmed fastq: reindeer --index --disk-query -f fof.unitigs.fa
• in the index directory, create a tsv (tab-separated) file named fos.txt. It must contain the list of the samples in the first column and the normalized value of the sample (computed using the kmer number of the fastq files of the sample) (needed by rdeer-service). Follow the link below to rdeer-service for more details.
• If you have used the --disk-query Reindeer index option in the directory index, create an empty file named disk-query (needed by rdeer-service).

software and python environment

Applications below are needed:

• kmerator https://github.com/Transipedia/kmerator
• rdeer-service https://github.com/Bio2M/rdeer-service
• snakemake https://github.com/snakemake/snakemake

In addition, certain python scripts need external libraries:

• PyYAML
• pyfaidx
• pandas
• plotly
• seaborn

All these are listed in the requirements.txt file, so, you can install all python software and libraries with pip:

pip install -r requirements.txt

R environment

The R script complexity.R needs the library Biostring, see documentation at https://bioconductor.org/packages/release/bioc/html/Biostrings.html

Loading the CCLE index in memory

Rdeer-service loads the index in memory and enables running multiple queries against the index in real time. Rdeer-service embeds a special version of Reindeer capable of running in socket mode.

We recommand taking a look at the README.md of rdeer-service to learn how to load the CCLE index: https://github.com/Bio2M/rdeer-service.

2. Quantification analysis

The quantification folder holds scripts for evaluating Reindeer's quantification of gene expression. See quantification/README for more detail.

3. Mutation analysis

Mutations analysis is provided as a Snakemake pipeline.

See mutations/README for detail.

4. Fusion / chimera analysis

The fusion/chimera analysis is computed with a bash script. See chimera/README for detail.

5. Transposable element analysis

See transposable_element/README for detail