search

darelab2014 / Dinopore

DInoPORE: Direct detection of INOsines in native RNA with nanoPORE sequencing
16 stars 3 forks source link

ground truth file #2

Closed ssscj closed 1 year ago

ssscj commented 2 years ago

Hi, thanks for developing Dinopore. I want the train the Dinopore model from scratch. And I found the groundtruth_class_regression.tsv and the groundtruth_class_regression_all.tsv in codeocean. And these two ground truth file only contain 3 columns, id, ref, rate, not the 5 columns mentioned in the README. Example file in misc on github is correct. Could you provide the whole ground truth file? Thanks!

acarmas1 commented 1 year ago

hey were you able to find the right .tsv file? I'm confused on how I'm supposed to generate this file

skchronicles commented 1 year ago

@hengjwj @darelab2014 @tramanh183 Is there an update to this issue?

I am also interested in using this tool with my own data; however, it is not clear what I should be provided to the -c option of the mainscript1.sh and mainscript1.sh. Just for reference, I am working with human data.

Please let me know what you think.

hengjwj commented 1 year ago

Hi all, I've removed the requirement for -c in mainscript1.sh as the subscripts do not require the file (mainscript1 and mainscript2 were originally together). As mentioned in #3, the file is used to assess model accuracy/precision based on ground truth from Illumina sequencing data. Kindly refer to #3 for the example files on formatting. Thanks!