LightStringGraph

Lightweight String Graph Construction

At this stage this software is a prototype/proof of concept.

If you want to try out software follow this steps:

1. download and compile BEETL (tested with version 0.4.0, set to 1 the flag BUILD_SA in src/shared/Tools.hh)
2. given your FASTA file a.fa create a new FASTA file b.fa that contains the reversed reads of a.fa (if read x is at position N in a.fa, read rev(x) should be in position N in b.fa)
3. build the BWT of a.fa beetl bwt -i a.fa -o BWT -f ascii
4. build the BWT of b.fa beetl bwt -i b.fa -o revBWT -f ascii
5. make all LighStringGraph
6. rebuild GSA[ $ ] ./bin/bgsa -i BWT.pairSA -o newSA.pairSA
7. run LightStringGraph ./bin/stringGraph -B BWT-B0 -R revBWT-B0 -G newSA.pairSA

stringGraph outputs (STDOUT) a json-like structure of "edges". Every edge has 3 fields:

1. from: source read ID (read position in a.fa)
2. to: destination read ID (read position in a.fa)
3. length: number of bases in the second read (the one pointed by to) that aren't in overlap with the first read

At this stage LightStringGraph works only with simulated genomic reads from the same strand and with only forward orientation. As stated before this is a proof of concept and will be improved later on.