rMATS-STAT: The Stand-Alone rMATS Statistical Model

This stand-alone rMATS code estimates the P values of differnetial alternative splicing using read counts on the alternative splcing events.

Requirements:

1. Install Python 2.6.x or Python 2.7.x and corresponding versions of NumPy and SciPy.
2. Add the Python directory to the $PATH environment variable.

Installation:

The source code can be directly called from Python.

Usage:

Use on replicates without pairs:

$ python rMATS_unpaired.py input_read_count_file output_folder number_processor diff_cutoff

Use on replicates with pairs (Each replicate is paired with another between the two sample groups):

$ python rMATS_paired.py input_read_count_file output_folder number_processor diff_cutoff

Use on pooled data (In each sample group, all the replicate read counts are added together):

$ python MATS_LRT.py input_read_count_file output_folder number_processor diff_cutoff

The 1st-2nd parameters specify the input and output. The 3rd parameter specifies the number of processors to run the code. The 4th parameter is the cutoff for splicing difference. The examples of input files are available in the depository.

Example:

Test rMATS statistical model on sample input: Run rMATS_unpaired.py, rMATS_paired.py and MATS_LRT.py as below to test the script.

$ python rMATS_unpaired.py inc.txt ./ 1 0.1
$ python rMATS_paired.py inc.txt ./ 1 0.1
$ python MATS_LRT.py inc.txt ./ 1 0.1

Input: One input file is required for the script. inc.txt: Each row of this input file contains an alternative splicing event. The 5 columns of this input file contain the read counts for the two isoforms of the alternative splicing event in the two sample groups (2-3 columns are for the first group, 4-5 columns are for the second group). The read counts for different patients are separated by commas in the column. As an example, for exon skipping events, each row defines a skipped exon and the columns contain the read counts for inclusion and skipping isoforms:

• ID: User defined ID for the alternative splicing event.
• IJC1: inclusion junction counts for group 1, patients are separated by comma.
• SJC1: skipping junction counts for group 1, patients are separated by comma.
• IJC2: inclusion junction counts for group 2, patients are separated by comma.
• SJC2: skipping junction counts for group 2, patients are separated by comma.
• IncFormLen: length of inclusion form, used for normalization.
• SkipFormLen: length of skipping form, used for normalization.

Output: The output folder contains the rMATS_Result_P.txt file. For each alternative splicing event, rMATS outputs the P-values that evaluate the associations between alternative splicing and patient survival.

• ID: User defined ID for the alternative splicing event.
• IJC1: inclusion junction counts for group 1, patients are separated by comma.
• SJC1: skipping junction counts for group 1, patients are separated by comma.
• IJC2: inclusion junction counts for group 2, patients are separated by comma.
• SJC2: skipping junction counts for group 2, patients are separated by comma.
• IncFormLen: length of inclusion form, used for normalization.
• SkipFormLen: length of skipping form, used for normalization.
• PValue: P-values of the alternative splicing event.

---

---

The following part covers the usage of a simulation code to generate simulation counts with an outlier in the samples. The simulaiton code requires the rpy module for Python.

Simulation Code Usage:

$ python Simu_Outlier.py total_read_count_file standard_deviation psi_min psi_max standard_deviation_outlier > output

The 1st parameter specifies the total read counts previously sampled from real data. The 2nd parameter specifies the standard deviation of psi without the sample group. The 3rd and 4th parameters specify the range of the mean splicing values. The 5th parameter specifies the standard deviation of the outlier. To simulate without outliers, please set the two standard_deviation parameters to be equal.

Example:

$ python Simu_Outlier.py count.txt 0.1 0 0.05 0.4 > output.txt

This example runs the simulation code with the total counts sampled in count.txt (available in depostory); with 0.1 standard-deviation in the sample group; mean psi values between 0% to 5% and 0.4 standard-deviation for the outlier.

---

---

The statistical code provided in the first section outputs the P values without multiple comparison correction. The following part provides a Python code to calculate the false discovery rate (FDR) based on the Benjamini-Hochberg approach.

FDR Code Usage:

$ python FDR.py input > output

The input file is the rMATS-STAT output from the previous step. The FDR.py code calculates the FDR and adds the FDR values to the last column of the output file.

Contacts and bug reports

Shihao Shen [email protected]

If you found a bug or mistake in this project, we would like to know about it. Before you send us the bug report though, please check the following:

1. Are you using the latest version? The bug you found may already have been fixed.
2. Check that your input is in the correct format and you have selected the correct options.
3. Please reduce your input to the smallest possible size that still produces the bug; we will need your input data to reproduce the problem, and the smaller you can make it, the easier it will be.

Publication:

Copyright and License Information

Copyright (C) 2015 University of California, Los Angeles (UCLA) Shihao Shen and Yi Xing

Authors: Shihao Shen and Yi Xing

This program is licensed with commercial restriction use license. Please see the attached LICENSE file for details.