Error rate / haplotype collapsing

ruanjue / wtdbg2

Redbean: A fuzzy Bruijn graph approach to long noisy reads assembly

GNU General Public License v3.0

513 stars 94 forks source link

Error rate / haplotype collapsing #40

Closed gbdias closed 6 years ago

gbdias commented 6 years ago

Hi Jue,

wtdbg2 produces very interesting results. I am particularly impressed with the low level of duplication (haplotigs) in the final assemblies.
Is there a combination of parameters that could reduce haplotype and repeat collapsing to produce an assembly with both alleles when you have heterozygous regions or structural variants? I am trying to replicate what Canu does, which is separating haplotypes that are 1-2% divergent.

Best, Guilherme

ruanjue commented 6 years ago

Thanks Guilherme,

When wtdbg2 assembles raw long noisy reads, it is not sensitive with the sequence similarity, so that haplotigs will be collapsed into unique ones.

Recently, I add a option '-s ' to filter alignments of accurate long reads from CANU-correction or PacBio CCS. Please note that 1-2% sequence divergence is not the same value in kmer-divergence, maybe 10% or others, need to calculations or tests.

Best, Jue

gbdias commented 6 years ago

Got it. Thanks for the info Jue.

Best, Guilherme