Arabidopsis VIRC instead of col0 (wild type)

GoekeLab / m6anet

Detection of m6A from direct RNA-Seq data

https://m6anet.readthedocs.io/

MIT License

100 stars 17 forks source link

Arabidopsis VIRC instead of col0 (wild type) #166

Open haotianteng opened 1 month ago

haotianteng commented 1 month ago

I am wondering if there is any reason that you trained and compared the model using the virC dataset (rescued) instead of col0 (wildtype) dataset for Arabidopsis?

yuukiiwa commented 1 month ago

Hi @haotianteng,

It was a while back, so I think we used virC instead of col0 due to public data availability.

Thanks!

Best wishes, Yuk Kei

haotianteng commented 1 week ago

Hi Yuk Kei, Thanks for the update. Also would you kindly provide which transcriptome and DNA reference genome was used to align the reads back?

haotianteng commented 1 week ago

Also I run the PRAUC curve from the virc_test_results.csv file downloading from here: https://codeocean.com/capsule/4723237/tree/v1

I got much higher PRAUC score (see below) compared to the score reported in your supplementary table. What could be wrong here? I used the "modification_status" as the true label.

m6Anet-VIRC - 0.4841165021344414 m6Anet-HCT116 - 0.3620230645743336 m6Anet-HEK293T - 0.43574739187201345