Final project for computational genomics. If you view in github, it's a lot prietter :)

This is a Guided/Assisted assembler for genomic sequences using a dynamic bwt (FM Index) that iteratively updates a reference sequence in an honest attempt to grow contiguous regions of dna (contigs) that are exact or near matches. This README serves several purposes:

1. Inform those who may stumble upon this of what they are looking at.
2. Comply with final project guidelines which state we should have a README.
3. Inform all that this is not an optimized code base yet, the emphasis is on functionality and accuracy in contig recognition.
4. Inform those that decide to pull/fork etc that this code comes with no warranty of anything.

• Disa Mhembere
• Stephen Lee

Note all library deps were only tested on MAC OS X 10.8/9

• Python 2.7
• Numpy 1.6.2
• Scipy 0.13.1
• For plotting Matplotlib 1.1.1 (optional)
• For generation docs Epydoc 3.0.1 (optional)

 assembler.py 

To run a full assembly you should use the script assembler.py. The best way to figure out what flags you should run it with is to type python assembler.py -h. This will reveal all command line flags and a description of each one's purpose. Here is an example of the args we generally pass:

 python assembler.py _ _ -r 20 -t -s -c 20 -p 0.01 -m 0.9 -n 100 -C 0.6 -e 

To reproduce Figure 2. in the paper, where we compare dynamic BWT to a static BWT updating reference, we used

 python assembler.py _ _ -r 30 -t -s -c 20 -p 0.01 -m 0.9 -n NT_SIZE -C 0.5 -e -O1 

for the static and

 python assembler.py _ _ -r 30 -t -s -c 20 -p 0.01 -m 0.9 -n NT_SIZE -C 0.5 -e 

We iterate through NT_SIZE values in

[150, 250, 300, 350, 400, 450, 500]

Note: Relaxing the -m (accuracy required for completion) value will lead to quicker completion times.

To reproduce Images like Figure 3, 4 an example would be:

 python assembler.py _ _ -r 30 -t -s -c 20 -p 0.01 -m 0.75 -n 100 -C 0.5 -e -P -S -F p100 

This produces an image named p100.png in your current working directory.

The important args to note here are:

• The first two positinal (required) args i.e. _ and _ are simply stubs for a reference input and a target input. They are overridden by the -t flag which runs the script in test mode, where both target and reference are synthetically generated and printed to stdout before assembly. If you specify real file names and don't use the -t flag we assume both inputs are 1 line sequences like those found in the data directory. You will still need to specify -r, the read length so that we can split and create near-even coverage.

You never have to directly interact with any other script, but note following:

• Every script that is not just a data structure contains a runnable main that has a simple test function that will run by simply passing:

 python script_name.py 

• Here script_name.py can be substituted with:

  • bwt.py
  • dynamic_bwt.py
  • reference.py
  • target_reads.py
  • aligner.py

• There is a tests directory that should contain unit tests. The dir currently contains tdynamic_bwt.py which is run by passing:

 python tdynamic_bwt.py 

Enjoy yourself ... We sure did while writing this!