Project

This is a package to help with the analysis of bulk RNA sequencing data. This package is built on scanpy.

State

This project is currently under development.

Installation of this package in another project

To install this library in another Python project, execute simply:

pip install git+https://github.com/idiap/bulkanalysis.git

Scripts

genes_preprocessing

genes_preprocessing.py aims at performing basic filtering and normalization on bulk RNA sequencing data. To run it, udpate the file config/genes_preprocessing_template.yaml with the right paths. You can then run the script as follows:

python3 scripts/genes_preprocessing.py --config_file config/genes_preprocessing_template.yaml

The config file should contain:

• data_origin: Origin of the transcripts matrix. For now, the only supported option is kallisto_whole_transcriptome, meaning that the transcripts matrix must come from a kallisto quantification on a whole transcriptome. Other options might be supported in the future.
• df_counts_path: Path to the matrix of transcripts.
• gtf_file_no_focus: Original GTF used for the whole transcriptome quantification.
• gene_names_path: GENCODE genes names with gene symbols. Example: "Gencode_geneNames_hg38V44.txt"
• gene_info_path: Genes symbol with their information, in particular whether they are protein-coding. Example: "Homo_sapiens.gene_info"
• treatments: dictionnary with name of the treatments in keys and list of corresponding sample names in keys.
• path_to_results: Directory where to save the results.
• figures_extension: Extension you want to save your figures with, e.g "pdf", "png",...
• pct_in_treatment: Percentage of samples within a treatment group in which a gene should be reliably expressed to be kept.

aggregate_featureCounts_output

aggregate_featureCounts_output.py aims at merging the outputs from featureCounts for multiple samples. To run it, run the script as follows:

python3 scripts/aggregate_featureCounts_output.py -f sample1.txt sample2.txt sample3.txt -n sample_name1 sample_name2 sample_name3 -s df_counts.csv

with:

• sample1.txt sample2.txt sample3.txt being the output files of the function featureCounts
• sample_name1 sample_name2 sample_name3 being the names of the samples you want to appear in the final matrix
• df_counts.csv: name of the file where to save the final matrix.