Code for Paired de Bruijn graph , based on recomb 2011 paper.  There has been some development since the paper, but you can still find the original code in Recomb2011version.zip

MAIN FILES:

singleGenerateReads.cpp  
Generates non-paired reads from a genome

generateReads.cpp  
Generates paired reads from a genome

ab.cpp       
Assembles non-paired reads

pab.cpp            
Assembles paired reads -- exact d version

pabNL.cpp    
Assembles paired reads -- NoLimit on delta

pabNLpar.cpp  
Threaded version of pabNL

pabFull.cpp  
Adds support for gaps in the coverage to pabNL.  This is done by effectively throwing away any matepairs whose right mate doesn't have a place in the DB (b/c the DB is implemented just for the left reads).
This version should supercede all the other ones -- use it in place of pab, pabNL, pabNLpar.  There are probably bug fixes in this one that are not in the other ones.




VALIDATION:

plot.py       
Plots the contig lengths

qcheck.sh    
Uses blat to check accuracy of contigs

qcheck2.cpp  
Uses suffix array to check accuracy of contigs




OTHER:

generateKs.cpp
Takes a readl file (with reads) and generates a readk file (with kmers).

check_loops.cpp    
Reports any loops found in the dbcollapsed output file.

union.h
Implements union find data structure

buildsa.cpp    
Builds the suffix array for a genome

filter_genome.cpp  
Removes all non [atcgATGC] characters from the genome and uppercases everything

run
Collections of commands used to generate figures in paper




THE FOLLOWING ARE FILES STILL IN DEVELOPMENT:

short.cpp                
A replacement for singleGenerateReads that saves memory by using the genome sequence in sorting.

abshort.cpp  
A replacement for ab.cpp saves memory by using indices into the  genome sequence in place of the read strings.

validate.cpp
A tool that takes a partition info and, using the source genome sequence as a cheat, checks if the partitioning is safe.

FILE FORMATS:

readl : each line is a matepair, left read followed by right read, no space in between
readk : each line is a left kmer in the reads and its position in readl
readks : readk sorted by the kmers with only the position stored

Workflow from a readl file is to first run generateKs to create readk, then sort to create readks, then pabFull to run assembly