search

anialisiecka / ALIBI

MIT License
7 stars 0 forks source link

ALIBI is slow #4

Open subwaystation opened 1 year ago

subwaystation commented 1 year ago

Hi there @anialisiecka,

I am applying ALIBI to a chr20 human pangenome graph https://s3-us-west-2.amazonaws.com/human-pangenomics/pangenomes/scratch/2021_11_16_pggb_wgg.88/chroms/chr20.pan.fa.a2fb268.4030258.6a1ecc2.smooth.gfa.gz and I had to abort it after 623 hours, because it did not finish.

Here is the command line:

/usr/bin/time --verbose bash alibi.sh -i chr20.pan.fa.a2fb268.4030258.6a1ecc2.smooth.gfa

Am I using ALIBI correctly? What could be the reason it takes so long?

Thanks for any help.

anialisiecka commented 1 year ago

Hi,

what is the size of your graph? How many nodes/edges are there?

Best,

Ania

subwaystation commented 1 year ago

These are the stats:

length: 187070843
nodes: 2820570
edges: 3969999
paths: 824
steps: 235459828
anialisiecka commented 1 year ago

Which version of ALIBI are you using?

subwaystation commented 1 year ago

I am using commit 2e982ff5a0410e8b8ec590e0cc826060c457b155. Which should be current master.

anialisiecka commented 1 year ago

I am not sure yet why it is working so slow. I need to think about this for a while.

anialisiecka commented 1 year ago

Hi,

based on our previous tests, I would expect ALIBI to finish the work within 7-14 days. But it substantially depends on the size and structure of the graph you want to linearize. I am going to perform some tests using your data set and try to understand the issue.

Best,