Coral (c-core)

Overview

Coral is a tool for analyzing proteomics data. It provides functionalities for data preprocessing, normalization, imputation, and statistical analysis using the limma package in R.

Docker Image Build

Building Your Own Docker Image

Building your own Docker image is optional. You can use the pre-built image from Docker Hub.

docker build . -t coral -f dockerfile/Dockerfile

Running from Built Docker Image

docker run --rm -v "./data:/data" coral -u "/data/For_Curtain_Raw_PPM1H- PROTAC_TP.txt" -a "/data/annotation.txt" -o "/data/output.txt" -c "/data/comparison.txt" -x "T: Index,T: Gene"

Running from Docker Hub Image

docker run --rm -v "./data:/data" noatgnu/coral:0.0.1 -u "/data/For_Curtain_Raw_PPM1H- PROTAC_TP.txt" -a "/data/annotation.txt" -o "/data/output.txt" -c "/data/comparison.txt" -x "T: Index,T: Gene"

Installation

Pip Install

Install R and set the R_HOME environment variable to the R installation directory. Also, install the QFeatures package and its dependencies.

pip install ccore-coral

Running from Pip Install

coral -u "/data/For_Curtain_Raw_PPM1H- PROTAC_TP.txt" -a "/data/annotation.txt" -o "/data/output.txt" -c "/data/comparison.txt" -x "T: Index,T: Gene"

CLI Usage

usage: coral [-h] [-u unprocessed] [-a annotation] [-o output] [-c comparison] [-x index] [-f column_na_filter_threshold] [-r row_na_filter_threshold] [-i imputation_method] [-n normalization_method] [-g aggregation_method] [-t aggregation_column]

Options:
    -u unprocessed, --unprocessed unprocessed
                        Filepath to the unprocessed data file.
    -a annotation, --annotation annotation
                        Filepath to the annotation file.
    -o output, --output output
                        Filepath to the output file.
    -c comparison, --comparison comparison
                        Filepath to the comparison file.
    -x index, --index index
                        Column names to be used as index.
    -f column_na_filter_threshold, --column_na_filter_threshold column_na_filter_threshold
                        Threshold for column-wise NA filtering.
    -r row_na_filter_threshold, --row_na_filter_threshold row_na_filter_threshold
                        Threshold for row-wise NA filtering.
    -i imputation_method, --imputation_method imputation_method
                        Method for imputation.
    -n normalization_method, --normalization_method normalization_method
                        Method for normalization.
    -g aggregation_method, --aggregation_method aggregation_method
                        Method for aggregation.
    -t aggregation_column, --aggregation_column aggregation_column
                        Column name to be used for aggregation.

Usage as a Module

import pandas as pd
from coral.data import Coral

coral = Coral()
# Read in the unprocessed data
coral.load_unproccessed_file("data/For_Curtain_Raw_PPM1H- PROTAC_TP.txt")
# Add sample column names
coral.add_sample("...")
# Add condition or group names
coral.add_condition("...")
# Add sample group mapping
coral.add_condition_map("condition_name", "sample_name")
# Add comparison
coral.add_comparison("condition_A", "condition_B", "comparison_name")
# Add index columns
coral.index_columns = ["index_column_name"]
# Filter columns by NA
coral.filter_missing_columns(0.7)
# Create QFeatures object
coral.prepare()
# Filter rows by NA
coral.filter_missing_rows(0.7)
# Impute missing values
coral.impute("knn")
# Log2 transform
coral.log_transform()
# Aggregate features
coral.aggregate_features("new_feature_column")
# Normalize
coral.normalize()
# Prepare limma matrix
coral.prepare_for_limma()
# Run limma
results = []
for d in coral.run_limma():
    results.append(d)
if len(results) > 1:
    # Merge limma results
    results = pd.concat(results)
else:
    results = results[0]
# Write results
results.to_csv("output.txt", sep="\t", index=False)