High nucleotide sensitivity/specificity, zero gene level

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:

High nucleotide sensitivity/specificity, zero gene level

Post by katharina »

Originally posted in the old forum by AB on 27.08.2014 - 00:22
sensitivity/specificity poor on gene level
Hi there,
I am trying to train Augustus on the beet genome because this is the closest related reference genome to my plant. But, I think I am doing something wrong. The results to training with first 100 random genes, then 1000 random genes, have yielded high nucleotide level sensitivity (0.936) and specificity (0.823), low values for exon level (0.394 sensitivity, 0.345 specificity), but 0 for the gene level. Does anyone have any suggestions?

Code: Select all

genes.gb:5149 
genes.gb.test:1000 
genes.gb.train:4149
Evaluation of gene prediction *******
| sensitivity | specificity |
|
nucleotide level | 0.936 | 0.823 |
/
| #pred | #anno | | FP = false pos. | FN = false neg. | | |
           | total/ | total/ |   TP |--------------------|--------------------| sensitivity | specificity |
           | unique | unique |      | part | ovlp | wrng | part | ovlp | wrng |             |             |
           |
           |        |        |      |               1620 |               1310 |             |             |
exon level | 2473 | 2163 | 853 | ------------------ | ------------------ | 0.394 | 0.345 |
           |   2473 |   2163 |      |  777 |  234 |  609 |  781 |  282 |  247 |             |             |
/
transcript | #pred | #anno | TP | FP | FN | sensitivity | specificity |
|
gene level | 735 | 560 | 0 | 735 | 560 | 0 | 0 |
/
User avatar
katharina
Site Admin
Posts: 531
Joined: Wed Nov 18, 2015 6:14 pm
Location: Greifswald
Contact:

Re: High nucleotide sensitivity/specificity, zero gene level

Post by katharina »

by katharina on 01.12.2014 - 16:37
There are two ways to treat stop-codons: either they are inluded in the CDS (i.e. the CDS ends exactly where the stop codon ends) or they are not (i.e. the CDS ends exactly before the stop codon begins).
AUGUSTUS has a command line parameter --stopCodonExcludedFromCDS which can be set acccordingly. By default, AUGUSTUS checks in the species' parameter cfg file, whether this flag is set to true or false.
Check whether your genbank file has the stop codon excluded or included, and then set the command line flag accordingly.
Post Reply