search

tanghaibao / jcvi

Python library to facilitate genome assembly, annotation, and comparative genomics
BSD 2-Clause "Simplified" License
738 stars 187 forks source link

ALLMAPS: stuck at concorde step #59

Closed MichelMoser closed 7 years ago

MichelMoser commented 7 years ago

Hi, I try merging and anchoring scaffolds from two genetic maps with following command: 

python -m jcvi.assembly.allmaps path --cpus 8 F2_K.pein102.map_ABH.finalOUT.allmaps.bed Peinf.genome.v1.0.2.scaffolds.fasta

It starts off well but gets stuck at the concorde step of the 2 or 3rd linkage group. It was stalling for 6 hours before i killed it.

STDOUT

Current iteration 480: max_score=(91.0,)
Current iteration 500: max_score=(91.0,)
12:15:37 [allmaps] No fitness improvement: (91.0,). Exit EC.
12:15:37 [allmaps] Dropped 0 minor scaffolds
$chr3 (F7_W-9|F7_K-3) FINAL
Peex113Ctg18043+ Peex113Ctg18472+ Peex113Ctg17683? Peex113Ctg01645? Peex113Ctg18494? Peex113Ctg18321? Peex113Ctg01900? Peex113Ctg02107? Peex113Ctg18340? Peex113Ctg18117? Peex113Ctg18593? Peex113Ctg18607- Peex113Ctg18319? Peex113Ctg08631? Peex113Ctg18183? Peex113Ctg17822? Peex113Ctg09431? Peex113Ctg18181? Peex113Ctg18375? Peex113Ctg07830? Peex113Ctg07311? Peex113Ctg01239? Peex113Ctg18658- Peex113Ctg05975? Peex113Ctg06037? Peex113Ctg18157? Peex113Ctg18506? Peex113Ctg03290- Peex113Ctg18144- Peex113Ctg18604- Peex113Ctg18416- Peex113Ctg17778? Peex113Ctg09565- Peex113Ctg17775? Peex113Ctg18339? Peex113Ctg18371+ Peex113Ctg18325?
12:15:37 [allmaps] Working on F7_W-10|F7_K-2 ...
12:15:42 [allmaps] Graph size: |V|=288, |E|=68642.
12:16:09 [base] `tsp_work` not found. Creating new.
12:16:09 [tsp] TSP rescale: max_x=520.0, min_x=-10000, inf=20000, factor=100
12:16:10 [tsp] Write TSP instance to `tsp_work/data.tsp`
12:16:10 [base] concorde -s 666 -x -o tsp_work/data.sol tsp_work/data.tsp >/dev/null 2>&1

Any ideas what i can do about it? Is it an issue with solving the TSP, so concorde can not converge?

Thank you, michel

tanghaibao commented 7 years ago

@MichelMoser yes it may be the graph is complex enough so concorde could not converge. You can try --skipconcorde to kick the search down to the GA step.

MichelMoser commented 7 years ago

@tanghaibao thank you for the quick answer. I just skipped the concorde-step for the specific chromosome by terminating the concorde command. But maybe i should skip it for all chromosomes for consistency as you told me.