discoverydna / map_motif

Map frequency matrices on sequence alignments.

Map Motif

About

This repository houses code that maps TFBS onto alignments. The purpose is to identify where TFBS are in an alignment. There are currently three aspects of the code, located in /code directory.

1. the map+motif.py script that can be implemented in command line.
2. motif_scoring_and_extraction.ipynb that is a jupyter notebook of the script for interactive coding and visualization
3. The /D3_vis portion, which is a working project to visualize TFBS interactively in the browser.

map_motif.py

map_motif.py is a python script that has two inputs.

File Inputs:

1. alignment (fasta)
2. TFBS Position Frequency Matrix.

Arguments:

1. alignment fasta file
2. TFBS Position Frequency Matrix
3. optional -threshold score cutoff (outputs only scores greater than the specified threshold)

File Outputs: -.csv file that outputs found TFBSs at each position, if any, in alignment.

Output data frame includes:

• position
• score
• sequence entry
• raw_position (from each sequence entry)
• strand (which direction the motif was found)
• motif_found (sequence motif at each postion)

The output file will be saved in directory script was ran.

Example output file: map_motif-alignment.fa-motif.fm.csv

To Run

python map_motif.py alignment.fa motif.fm 3.2

map_motif's Issues

Read in only one raw sequence option

It would be cool to allow the program to run only one sequence. Maybe have the program print out the number of sequences it reads when running the program, so the user knows how may sequences it is running.

Make tools to convert PWM

or at least a guide to the cluster fuck of data types and other tools that kinda help with this

Line 92 Cannot be hardcoded

For some reason this is hardcoded in. Change this to be the length of how many sequences were input.

File input for Jupyter Notebook

The only way that the file input works for the Jupyter notebook implementation is that the files have to be in the same directory. See 'motif_mapping.ipynb`, first chunk.

We should have a variable that specifies 1. where the input files are and 2. where the output files should go.

set threshold

set threshold as an argument

Add visualization

Add in easy ggplot visualizations based on table output

For command line version: update new features.

For the command line version, update new features from ' Motif_Scoring_and_Extraction.ipynb'.

Including

1. PWM visualization output option (as argument)
2. Sequence Overview Visualization option (As argument)
3. Threshold as a percentile instead of score

clarification of line

I still don't know what this does

positions = {'score':score, 'position':position, 'species': i}
position_list.append(positions) 

line 136

map_motif.py table output to .csv

The file extension for the map_motif.py program say ".csv", but the file is tab separated. Either change the extension to .tsv or change the output to .csv file type.

a lot of the data frames have columns where the same piece of text is fills the whole column. Would it be more efficient if we stored it in one place instead of for every column in the data frame?

I still don't know what this does

positions = {'score':score, 'position':position, 'species': i}
position_list.append(positions) 

line 136

Optional Argument Proper Syntax?

In map_motif.py are we doing the most appropriate style for optional arguments?

See map_motif.py line 32

It would be great to make this a python package rather that a script that is run from command line.

Read through code and see how to make more efficient and more pythonic.

Since I am not a python programmer, I am not sure the most efficient and proper way to make this program.

Make map_motif.py script into Python 3 compatible

Redundant files and organization

The code folder is a little crazy. attn: @alexnakagawa, @jshi1010

Few issues:

1. What is the difference between these two files?
   motif_mapping.ipynb & Motif_Scoring_and_Extraction.ipynb. Is the motif_mapping file just for D3 building?

2. Motif_Scoring_and_Extraction.ipynb filename should be lowercase: motif_scoring_and_extraction.ipynb

3. We should split the command line code (map_motif.py) from the Jupyter notebook code and maybe even the an additional folder for D3 code. The structure could be py, ipynb, and D3. Thoughts?

Add print statements.

For the command line version. There are many things that need to be printed.

• Like make sure the user is inputting alignment files. Maybe just checking if the files have "-" and print out if the sequences are NOT alignments. Or just print how many files and what type of files are read in
• error messages (like "this is not a fasta file"
• success print statement and tell user where output file is located.