Write a separate tool or function to scan sequences, in order to avoid dependency to RSAT matrix-scan-quick C program

pvhelden / optimize-matrix-GA

Optimise one or several position-specific scoring matrices based on their capability to discriminate positive from negative sequence sets.

GNU Affero General Public License v3.0

0 stars 0 forks source link

Write a separate tool or function to scan sequences, in order to avoid dependency to RSAT matrix-scan-quick C program #39

Open jvanheld opened 1 month ago

jvanheld commented 1 month ago

The installation and use of omtimize-matrix-GA currently depends on RSAT for the sequence scanning tool matrix-scan-quick, implemented in C. This dependency imposes to have either an RSAT suite (e.g. Docker, apptainer) or to compile the C tool (ot teste yet).

We should evaluate the efficiency of a full-python solution.

jvanheld commented 1 month ago

Input

A python script that takes as input :

a sequence set (fasta format)
a position-count matrix (transfac format)

Output

2 options :

all the matches for the PCM
only the best match per sequence (option -best_match)

Arguments:

residue probabilities a dictionary with default : {A:0.25, C:0.25, G:0.25, T:0.25}
pseudo-count (default : 1)

jvanheld commented 1 month ago

Steps

[x] read matrix from transfac file
[ ] Compute the Position-Weight Matrix (PWM) from the PCM and the priors (see RSAT slides 02.3.pattern_matching_matrices.pdf for the formula)
[ ] read fasta sequence
[ ] scan a sequence with the PWM and prints matches