categorising the result

[dikshyantg]

Hi , running the Deepgo plus with the basic command gave me result.tsv file with predicted go terms for each sequence. However, I wanted to categorise those GO terms, based on cellularcomponent , molecular function and biological process for each fasta sequence. The tsv files , I get even for one gene has large number of predicted go terms with their statistical values separated by "|". However they aren't arranged in a descending order so predicting which is the better go term is little tricky so categorising them first then looking at the highest statistical value seems like a better way for me. Would you be able to help me in this ?

Thanks!

[coolmaksat]

Hi, thank you for raising this issue.
You can use predict.py script to get better formatted results. I've just added the GO term subontology to it.

[dikshyantg]

Thanks a lot that worked. Though unlike Deepgo command , this predict.py script requires some files under data folder as zipped so that gzip could open the file. Another difference was Deepgo gave the result in tsv like GO:0032502|0.318 which means the statistical value is 0.318 but here it is given in tsv as HRE53_03560 GO:0016787 molecular_function hydrolase activity 1.34 0.208 . believe it is saying between 1.34 to 0.4208 and we choose the later (0.4208)right?

[coolmaksat]

Hi,
yes, the last item is the prediction score. The number before is the class specificity (or information content).