search

mikolmogorov / Flye

De novo assembler for single molecule sequencing reads using repeat graphs
Other
789 stars 167 forks source link

running time for Flye-assembly #225

Closed yifeng-evo closed 4 years ago

yifeng-evo commented 4 years ago

I am assembling Nanopore raw reads using meta mode. The metagenome size is not clear so I used 200m. And I used 8 threads. It has been running for 3 days and I keep seeing the updates in log like this:

[2020-02-26 16:32:11] DEBUG: Ovlp index size: 86674963
[2020-02-26 16:32:11] DEBUG: Inner: 2880602 covered: 3205552 total: 3707242
[2020-02-26 16:36:58] DEBUG: Assembled disjointig 727
        With 5 reads
        Start read: +ce917c7c-7d48-4360-8661-dcd437600ece
        At position: 1
        leftTip: 1 rightTip: 1
        Suspicious: 1
        Mean extensions: 1
        Avg overlap len: 11202
        Min overlap len: 5809
        Inner reads: 0
        Length: 38710
[2020-02-26 16:36:58] DEBUG: Ovlp index size: 86721189
[2020-02-26 16:36:58] DEBUG: Inner: 2880618 covered: 3205732 total: 3707242
[2020-02-26 16:40:11] DEBUG: Assembled disjointig 728
        With 5 reads
        Start read: +f3e1191f-4d53-450c-8f0a-e27b9ef0bf6e
        At position: 1
        leftTip: 1 rightTip: 1
        Suspicious: 1
        Mean extensions: 3
        Avg overlap len: 32443
        Min overlap len: 9150
        Inner reads: 0
        Length: 40726
[2020-02-26 16:40:11] DEBUG: Ovlp index size: 86721748
[2020-02-26 16:40:11] DEBUG: Inner: 2880640 covered: 3205868 total: 3707242

But there's nothing in 00-assembly folder yet. Is it typical that this stage spends so much time? Thanks a lot for your help!

mikolmogorov commented 4 years ago

Depending on the genome size and coverage, some assemblies definitely take longer than 3 days. You can always use more threads to speed it up.