Prototype code for converting HAL to Side graph
Iteratatively add genomes to side graph. Each genome is aligned to nearest genome already in side graph. This alignment is used to thread the genome onto the graph.
Dependencies: HAL and SonLib. Expected to be in sister directories to hal2sg but can be changed in include.mk. Tested only with latest version from development branch for both.
Note Best to compile in debug mode (cppflags_dbg in sonLib/include.mk). asserts are used to check if the output is honest. At the very least run the unit tests:
make test
To run the converter:
hal2sg input.hal output.fa output.sql > output.out
input.hal Input alignment to convert
output.fa Output fasta file of all Side Graph sequences
output.sql Output text file listing INSERT commands for Sequences, Joins and Paths (for each input sequence) in the graph.
- None of the options except
--targetGenomes(and the standard HAL caching options) work. - Please ignore voluminous garbage printed to stdout.
- SQL output needs work. Obvious things are directly updating db; verifying use of references, variants etc.
hal2sg only works on input of this form:
- Mutation free: ie as produced by Adam's pipeline
- Root reference: root genome used as initial side graph sequence.
- Star Tree
- Cactus graphs. Need to finish mutation support
- Arbitrary topology and references and genome subselection. Much is done, just need to generalize path code, and handle paralogies in insertions, and of course tests.