Splicemod

A toolkit for scoring and modifying exons and their adjacent intronic boundaries.

Author: Daniel Bryan Goodman ([email protected])

Python Dependencies

We use Python 2.7.

Required python packages are in requirements.txt and so can be installed with:

pip install -r requirements.txt

Note that biopython MUST be version 1.57, which is quite old, as splicemod uses the deprecated motif package. It is recommended that you install the requirements in a python virtualenv.

Bash commands for set-up

Read below for more info, but this set of commands should set up your virtual environment, install the packages, and download the wiggle tracks:

#choose a virtual environment dir and set it up
venv_dir= ~/.pyenv/splicemod-venv
virtualenv $venv_dir
source $venv_dir/bin/activate

# from the splicemod dir:
pip install -r requirements.txt
source scripts/get_wig.sh

Running splicemod

After installing the required data (see below) and python packages, Splicemod can be run from the base directory with the command:

`python src/ensembl.py`

This will write gbk/fas files for the natural and mutated exons to the data/ccds_ensembl dir. It might also be useful to save the output log to a file, like:

`python src/ensembl.py > data/2017.02.23.splicemod_output.txt`

Data required

• Motif definitions are included in the data/motifs dir.

• Ensembl mySQL database access (local or remote, see below)

• Wiggle tracks for conservation. This requires approximately 5.3 Gb. These can be downloaded and indexed with an included bash script:

source scripts/get_wig.sh

Ensembl Database

Remote Ensembl Access

Ensembl can be used remotely and the host and port can be set in src/cfg.py. The defaults currently work correctly but a list of up-to-date urls can be found here.

Local Ensembl Copy

A local copy of the ensembl database can also be used for fast access. These directions are based off of the ENSEMBL guide found here:

http://useast.ensembl.org/info/docs/webcode/mirror/install/ensembl-data.html

Download the ENSEMBL sql files for both core and * from the Ensembl FTP site and unzip them:

mkdir -p /path/to/ensembl_db_dir/core
mkdir -p /path/to/ensembl_db_dir/variation
cd /path/to/ensembl_db_dir/core
wget -r ftp://ftp.ensembl.org/pub/release-78/mysql/homo_sapiens_core_78_38/
cd /path/to/ensembl_db_dir/variation
wget -r ftp://ftp.ensembl.org/pub/release-78/mysql/homo_sapiens_variation_78_38/
gunzip *.gz

Install mysql if not already installed, and create a DB in the ensembl console:

create database homo_sapiens_core_78_38;

Then load the schema. I created a user called ensembl and gave it full access to the new db.

mkdir -p /path/to/ensembl_db_dir/core
mysql -u ensembl homo_sapiens_core_78_38 < homo_sapiens_core_78_38.sql
mkdir -p /path/to/ensembl_db_dir/variation

Then load all the txt data into the new schema. This takes a while.

mysqlimport -u ensembl --fields_escaped_by=\\ homo_sapiens_core_78_38 -L *.txt

Cached Exon List

To speed things up, we cache a copy of all exons in the file data/ccds_ensembl/78_38_CCDS_exons.all.txt. This file was generated using the mySQL command in get_ccds_exons() function in ensembl.py and is hard coded with an exon size of 100. The easiest way to regenerate this file is to run the mySQL query in a program like Sequel Pro and copy the result into a text file, the filename of which is pointed to in cfg.ens_exon_fn.