[maker-devel] tradeoff between run time & file number

[Daniel Ence]

Hi Rebecca, So, as far as pruning down the dataset goes, I think that the biggest gains will be made by trimming the number of scaffolds that you annotate. What is the n50 of your 400,000 scaffold set? Usually, scaffolds shorter than 5k or 10kbp won't contribute much to the gene counts in the end.

Also, if you can, try to avoid using the alt_est option. It works completely fine, but blasting those sequences takes much longer than blastn or blastp.

Otherwise, I'd need to see your maker_opts.ctl file to see how you've got things set up. You can attach those to your reply (to the maker-devel list), and I'll take a look. I don't how to force maker to create fewer files. You definitely want to be able to make use of the results from prior runs to save time.

Thanks,
Daniel

Daniel Ence
Graduate Student
Eccles Institute of Human Genetics
University of Utah
15 North 2030 East, Room 2100
Salt Lake City, UT 84112-5330
________________________________
From: maker-devel [maker-devel-bounces at yandell-lab.org] on behalf of Rebecca Harris [rbharris at uw.edu]
Sent: Wednesday, March 19, 2014 7:19 PM
To: maker-devel at yandell-lab.org
Subject: [maker-devel] tradeoff between run time & file number

Hi -

I'm running maker on a dataset of >400,000 scaffolds with MPI -n 64. I've gone through it once - and used the clean_up option because otherwise maker exceeds the clusters file_quote. However, now I'm retraining SNAP and it is taking a very long time - probably because it has to go through BLAST again. Is there anyway of getting around this? I expect I may have to train SNAP and rerun maker multiple times and it is taking about 3 weeks to get through my dataset. Is there a way to prune down my original dataset based on maker's output?

Thanks,
Rebecca
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20140320/c636afd0/attachment-0003.html>