reduce_redundancy

Purpose

The goal of this script is to reduce the redundancy in a set of sequences.

Input

• A FASTA file of sequences. FASTA entry descriptions are limited to less than 128 characters.
• Distance threshold.

Process

Iterate through each transcript in the dictionary and:

1. Generate a distance matrix from the sequences in the input file using Clustal Omega.
2. Parse distance matrix to dictionary.
3. For each FASTA entry, identify as a cluster those other sequences with a distance less than given threshold.
4. For each sequence in each cluster, generate a total combined distance score to the other members in the cluster.
5. Keep the entry with the lowest distance to the other members of the cluster as a representative sequence. Ignore the other members of the cluster.
6. Keep all other entries which have not included in any cluster.

Output

• A FASTA file of all cluster representative sequences and non-clustered sequences.

Dependencies

• Currently only tested on Linux
• Python 2.7
• Biopython (http://biopython.org/wiki/Download#Installation_Instructions)
• clustalo sudo apt-get install clustalo

Usage

• Edit the input_seqs_filename variable of the 'Initiating Variables' section to point to the input sequences.
• 12,610 related protein sequences (retrieved by iterative BLAST search) reduced to 2,048 sequences, using a distance threshold of 0.10, in 20 minutes on an i7 processor.