negative false problem

[mujiezhang]

Hi,
I performed inovirus_detector in Genbank viral genomes, but I found some typical inovirus like M13 and f1 could not be predicted by it. And a total of 53 inovirus were missed among all 338 inovirus in NCBI. Do you have any suggestions to improve the accuracy of prediction? It is strange that M13 and f1 were in your reference dataset but they could not be predicted by inovirus_detector.

[simroux]

Hi,

There is no active development on this project, so we won't have any new feature or capability to improve the accuracy of prediction. However, it is a little strange that it is missing M13 and f1. One thing you could double-check what score you get for these sequences, by changing the script "Get_inovirus_prediction_score_from_gff_fragments.pl": on line 11, the default threshold is 0.83. You may want to change this number (e.g. to 0.5), and rerun the pipeline on your test inoviruses, to see if maybe these are detected but with a lower score ?

[mujiezhang]

Thanks for your suggestion