past approaches they have tried, and why they failed/what status was

PBJelly + Quiver: gap filling MHAP -> low coverage after error correcting

current approach you're trying

PBcR-MHAP

what problems you are running into, or that you anticipate running into

PBcR-MHAP is based on wgs-assembler (Celera). They only support SGE and LSF. Submitting a single big job consumes too much memory (gatekeeper -> 1.5TB) or could use more CPU (overlapper) on some steps

what other approaches people are using

Josh: I just worked on a fungus genome which we used PacBio (Titus is involved with this). I used PBJelly and then SPAdes and got a really good assembly when paired with our Illumina data from the original sequencing project (PacBio was for genome scaffolding improvement).

Eichler lab: HGAP (using a modified Blasr, but they just added some outputs, no core modifications)

Dazzler - http://dazzlerblog.wordpress.com/ Currently just the aligner, need to figure out how to use together with other parts of the pipeline. Talk with Jason Chin, he done that for big genomes

ectools - https://github.com/jgurtowski/ectools

dbg2olc - http://arxiv.org/abs/1410.2801 http://sites.google.com/site/dbg2olc/. claim to be superfast

http://wgs-assembler.sourceforge.net/wiki/index.php/RunCA

Adam tips

Hybrid

dbg2olc

• arxiv
• u. mariland

lordec

ec2tools

pure

pacbio2ca

• Align with reference (Blasr, galGal4)
• Smaller subsets:
  • Unmapped reads
  • Reads mapping partially to micro chromosomes