Experiment with major assembly polishers about long-read-pipelines HOT 7 CLOSED

I'm interested in trying out MarginPolish so I'll add that as well and close this issue.

from long-read-pipelines.

How is "interesting" defined?

from long-read-pipelines.

• GPU acceleration seems slow down one step slightly, but significantly speeds up another step.
• The resulting assembly was polished to be quite a bit longer than the original, suggesting many indel errors in the original assembly, assuming the polish made the assembly better in indel rate.

from long-read-pipelines.

@ericsong @kvg should this be closed (or replaced with another ticket) given recent PRs?

from long-read-pipelines.

Are there others we'd like to add? Right now, we have Nanopolish, Racon, and Medaka.

from long-read-pipelines.

That's all that I know of.
Any others, @kvg ?

from long-read-pipelines.

We've found satisfactory answers to this on malaria data, but we should be mindful that different recipes have difference performances on different genomes. We'll likely need to evaluate polishing procedures quite often for different organisms.

from long-read-pipelines.