PyDESeq

Introduction

PyDESeq is a package to process control and treatment files to calculate differential gene expression. It calculates the log2 fold change and p-values for each gene and generates a filtered volcano plot. The script requires control and treatment FPKM (Fragments Per Kilobase Million) files and a gene name file.

Prerequisites | Install Instructions | Basic Usage | Input file format | Output files | Example and testing | Contributors | Supplemental information

Prerequisites

pydeseq requires:

• Python 3.xx
• Python packages:
  • pandas
  • numpy
  • scipy
  • matplotlib

If these packages are not yet installed, use the pip command:

pip install pandas numpy scipy matplotlib

If you do not have root access, you can run the command above with the additional --user option to install locally:

pip install --user pandas numpy scipy matplotlib

Install Instructions

pydeseq can be installed with the following commands:

git clone https://github.com/shaniabuu/PyDESeq.git
cd PyDESeq
pip install -e .

If you do not have root access, you can run the pip install command above with an additional option to install locally:

pip install --user -e .

To run pydeseq sucessfully, your working directory should be in the PyDESeq directory, otherwise, you might need to add this directory to your PATH. Note, this will have to be run again after opening a new terminal:

export PATH=$PATH:~/.local/bin

Typing pydeseq --help should show a useful message and can be run to see if the install was successfull.

Basic Usage

The basic usage of pydeseq to process control and treatment files to calculate differential gene expression is:

pydeseq [-h] -c {CONTROLS ...} -t {TREATMENTS ...} -g {GENE_NAME_FILE} -p {PVALUE_THRESHOLD} [-o {OUTPUT_DIR}]

pydeseq options

The required inputs for pydeseq are:

• -c, --controls: Control file. See Input File Formats for more details.
• -t, --treatments: Treatment file. See Input File Formats for more details.
• -g, --gene_name_file: Gene name file, for converting gene IDs to gene names. See Input File Formats for more details.
• -p, --pvalue_threshold: Sets p-value threshold for the output of the number of differentially expressed genes and the volcano plot.

Other additional options are:

• -h, --help: Shows the help message and exits
• -o, --output_dir: Sets output directy for output files. If not directory is given, saves output files to current working directory.

Input file formats

Control and Treatment files

Control and Treatment files that are used should contain just two tab-delimited columns, the first for gene ids and the second for FPKM values:

gene_id      FPKM

• column 1: gene_id stores the gene ids for the sequenced genes.

• column 2: FPKM stores the FPKM (fragments per kilobase of transcript per million mapped reads) values of the sequenced genes.

Note: There should be two or more files for control and treatment files as replicates for the calculation of p-values.

Note: The file should contain a header for proper use.

See text files in /data/lab_data for example.

Gene name file

The gene name file that is used should contain just two tab-delimited columns, the first for gene ids and the second for gene names:

• column 1: gene_id that includes those in control and treatment files.

• column 2: gene_names for the readable gene names for each gene id.

Note: the file should not contain a header for proper use.

See GRCm38.75.gene_names in /data/lab_data for example.

Output files

text outputs

1. Top 10 Genes with the Smallest Non-zero P-values: The script prints the top 10 genes with the smallest non-zero p-values, showing their gene names, log2 fold changes, and p-values.
2. Number of Differentially Expressed Genes: The script prints the number of differentially expressed genes after filtering with the provided p-value threshold.

output files

differential_expression_results.csv

A comma-separated value (.csv) file with the differential expression analysis results stored in the following columns:

gene_id      log2FoldChange    pvalue

column 1: gene_id stores the gene IDs for the identified differentially expressed genes.

column 2: log2FoldChange stores the values of the log2 fold change in expression of the gene between control(s) and treatment(s).

column 3: p-value stores the p-value for each gene where the null hypothesis is that that the gene expression is the same in control(s) vs. treatment(s).

volcano_plot.png

This image file is a volcano plot of the log2 fold changes vs. -log10(p-values). Genes with p-values below the threshold are highlighted in red. The top 10 genes with the smallest non-zero p-values are labeled.

See Supplemental information for more details about how the outputs are generated.

Example and testing

• To test the package using lab data:

cd PyDESeq
pydeseq -c data/lab_data/Chow_Rep1.txt data/lab_data/Chow_Rep2.txt  data/lab_data/Chow_Rep3.txt \
  -t data/lab_data/HFD_Rep1.txt data/lab_data/HFD_Rep2.txt data/lab_data/HFD_Rep3.txt \
  -g data/lab_data/GRCm38.75.gene_names  -p 0.05 -o test/lab_data/

The input files include three replices of mouse with standard "chow" diet as control and three replicates of mouse with high fat diet ("HFD") as treatment. The gene_id and gene_name conversion file is GRCm38.75.gene_names stored in the same directory as input files. The p-value threshold is set at 0.05 and the output file directory is test/lab_data/.

See /test/lab_data for output with this command.

• To test the package using public data:

cd PyDESeq
pydeseq -c data/public_data/DOXY+_FUS-WT_REP1.txt data/public_data/DOXY+_FUS-WT_REP2.txt \
  -t data/public_data/DOXY+_FUS-WT_SGenr_REP1.txt data/public_data/DOXY+_FUS-WT_SGenr_REP2.txt \
  -g data/public_data/Human_gene_names.txt  -p 0.05 -o test/public_WT/

cd PyDESeq
pydeseq -c data/public_data/DOXY+_FUS-P525L_REP1.txt data/public_data/DOXY+_FUS-P525L_REP2.txt \
  -t data/public_data/DOXY+_FUS-P525L_SGenr_REP1.txt data/public_data/DOXY+_FUS-P525L_SGenr_REP2.txt \
  -g data/public_data/Human_gene_names.txt  -p 0.05 -o test/public_FUS/

Contributors

This repositiory was generated by Shania Bu, Amber Tse, and Janice Wu for our S24 CSE 185 final project, inspired by the R package DESeq2.

Supplemental information

Package Workflow

The figure illustrates the program's logic. Black circles denote inputs, rectangles indicate outputs, blue text specifies the input file formats, and red warning text highlights where error messages are generated within the program.