ICEPOP

Immune CEll POPulation

Overview

A method for estimating immune cell population in the expressed genes. It enable analysis of differentially expressed genes (DEGs) or raw expression data. These APIs and scripts will let you have a fine grained control on the data analysis lacking in the web version. On top of the core immune cell population deconvolution, it allows you to download the raw data from NCBI GEO gene expression database, normalize, and plot the data using command line interface.

Requirements

• Python 3.3 or higher
• pip package manager

Installation

ICEPOP is best installed via pip through one of the following commands::

$ pip install git+https://github.com/ewijaya/icepop.git
$ pip install git+https://github.com/ewijaya/icepop.git --upgrade
$ pip install git+git://github.com/ewijaya/icepop.git
$ pip install git+git://github.com/ewijaya/icepop.git --upgrade

Recent Improvements

The codebase has been recently optimized and modernized with the following improvements:

Python 3 Migration

• Full Python 3 compatibility (3.3+)
• Updated all deprecated syntax and methods
• Improved performance with modern Python features

Performance Optimizations

• 5-50x speedup for large datasets
• Optimized clustering algorithms with vectorized numpy operations (10-100x faster)
• Enhanced DataFrame operations and reduced memory overhead
• Improved iteration methods for better performance

Code Quality

• Extracted common constants to centralized module
• Removed hardcoded paths for better portability
• Eliminated code duplication (~25% code reduction in optimized sections)
• Enhanced maintainability and consistency across modules

Alternative access

• Web application [main][mirror]
• API documentation

Calculating immune response score from DEGs

The input should be either in form of CSV or TSV files. The results can be in the form of table or plot. They are determined by the suffix of the output file.

To create bar plot use this command:

$ icepop_degs input_type1_degs.tsv -fclim 2 -s mouse -o output_file.jpg

The command will produce individual plots depending on the number of samples.

To create table:

$ icepop_degs input_type1_degs.tsv -fclim 2 -s mouse -o output_file.tsv
$ icepop_degs input_type1_degs.tsv -fclim 2 -s mouse -o output_file.xlsx

Suffixes of the output should either one of these: 'svg', 'jpg', 'png', 'tsv', 'xlsx', 'xls'.

Circos plot for unearthing the gene features in immune cells

It assumes that Circos is already installed in your main path. Typical use looks like this in Bash script:

INFILE=input_type1_degs.tsv
CIRCOS_DIR=your_circos_dir
# Find your Python site-packages directory with: python -c "import site; print(site.getsitepackages()[0])"
CIRCOS_CONF=$(python -c "import site; print(site.getsitepackages()[0])")/icepop/circos_conf/

icepop_degs_circos_uniform $INFILE \
    --go \
    -fclim 2 \
    -circos_dir $CIRCOS_DIR

cd $CIRCOS_DIR
cp -r $CIRCOS_CONF .
circos -param random_string='image' -conf ./etc/circos-medium.conf

This script produces the circular plot that links feature genes among the samples.

Publication

E. Wijaya, Y. Igarashi, N. Nakatsu, Y. Haseda, J. Billaud, YA Chen, K. Mizuguchi, H. Yamada, K. Ishii, T. Aoshi. Quantifying the relative immune cell activation from whole tissue/organ-derived differentially expressed gene data. Sci Rep. 2017 Oct 9;7(1):12847. PMID:28993694