Closed kvittingseerup closed 2 years ago
Hi @kvittingseerup,
Indeed ;)
We are working on it :) Hope to get back to you soon about it.
What we have used up until now is the pipeline
isONclust -> isONcorrect -> IsoCon
where we have run IsoCon
on each corrected-reads cluster from isONcorrect
. We then set the parameter --ignore_ends_len 30
to IsoCon
(default 15) to account for larger differences in 3' and 5' ends that what IsoCon
was designed for. IsoCon
was originally designed for targeted data using start and end primers where the 3' and 5' sites were fairly fixed.
It is definitely not a perfect solution both from an accuracy and a runtime aspect, in large part due to the variability in (ONT) data at 3' and 5' ends. The problem needs a tailored solution.
Best, Kristoffer
Hi Kristoffer
Thanks for the quick and thorough response!
Impressive work btw!
Cheers Kristoffer
Hi Kristoffer (great name btw ;) )
Do you know if it is possible to make a reference-free transcript assembly from the output of isOncorrect?
Cheers Kristoffer