Closed mp15 closed 3 years ago
Yes, source code in GitHub is totally out of date. But we can still take your change locally.
I think we can simply call this job.step.dust
, since DBdust is the main work. Actually building the database is pretty quick. We might split DBdust into its own parallel tasks someday, unless we make it multithreaded. Either way it deserves its own section.
I've pulled your branch, and I will merge your change, with job.step.dust
as the section-name. It will be in bioconda in a week or so, I think.
re: https://github.com/PacificBiosciences/pbbioconda/issues/28
Thanks for taking this, I've hacked my python library as an interim solution.
Please feedback that the loss of Falcon as a properly open source product utterly sucks for those of us that use customisations to optimise their assembly pipeline (or even just make it work in time for a deadline). It would probably be helpful if you created a set of instructions for a supported method to patch your conda recipes for those of us who want to take the training wheels off.
Just hit this again in second dust step in the 1-preads_ovl/build so it might be a good idea to make that configurable under the same job.step.dust configuration item please.
The build and dust step is currently scheduled with a memory requirement of 4000MB. This appears to be insufficient for dusting an 80x human. Make this option configurable from the config file.
I have had difficulty testing this with current FALCON as git seems to be out of date.