Specificity of augustus

Discussions about training AUGUSTUS from various sources of evidence. Not discussed here: BRAKER1 and WebAUGUSTUS!

Moderator: bioinf

Post Reply
User avatar
katharina
Site Admin
Posts: 531
Joined: Wed Nov 18, 2015 6:14 pm
Location: Greifswald
Contact:

Specificity of augustus

Post by katharina »

Originally posted in the old forum by Song on 04.04.2014 - 09:40
Hi,
I am try to retraining augustus with RNA-seq data.
Firstly, I took the TAIR10 annotation itself as trianing dataset, and then predict the gene structure of Arabidopsis col. And got a result 1.
Secondly, I used the PASA pipeline to generate a dataset with RNA-seq as training data. And predict the gene structure of Arabidopsis col. Got a result 2.
And found the specificity(true positive/predicted positive) in transcript level of both results are about 60%. So I think the training dataset is good enough, is it correct?
But he specificity(60%) is still far from enough. So I am wondering what's the general value of specificity. And how could I improve it.
User avatar
katharina
Site Admin
Posts: 531
Joined: Wed Nov 18, 2015 6:14 pm
Location: Greifswald
Contact:

Re: Specificity of augustus

Post by katharina »

by katharina on 07.04.2014 - 11:33
A specificity of 60% on transcript levels seems rather good. Have a look at the accuracy values of RGASP: http://www.nature.com/nmeth/journal/v10 ... .2714.html
Post Reply