Chris Wright
Chris Wright
I don't think it's reasonable to assume that memory errors in another piece of software imply a memory error in medaka. To test that hypothesis it would be straightforward enough...
It would help greatly if you could create a fork of medaka and push a branch with the changes you have made.
@gordonrix, @rob234king I've spent some time looking at this today. @ftostevin-ont previously looked at the issue and could only reproduce the error whilst using a GPU for the medaka RNN...
It may well be that we have multiple issues with the same sympton. What @ftostevin-ont and I have both shown is that you can take the intermediate outputs from `medaka...
Medaka v1.7.0 contains the fixes I implemented yesterday, and is now available on PyPI.
Hi @edgardomortiz, This is a nice suggestion and something we've thought of doing in the past. Unfortunately while relatively straight-forward this is something we don't have the free development time...
You'll have to take up the issue with the conda package with the people at @bioconda I'm afraid; I've given up on that front, see my comments [here](https://github.com/bioconda/bioconda-recipes/pull/33669#issuecomment-1069285343). Installation with...
Apologies, this is not currently possibly. You will have to write a bam without read group information or with a single read group. I cannot unfortunately give a timeline on...
We do not have generic automated scripting that implements what is noted in the README. It wouldn't be too difficult to add to the existing `medaka_consensus` script in a way...
The core algorithm of medaka is applied to overlapping chunks of base counts matrices derived from read pileups (much like samtools tview would show). These chunks are stacked to form...