tags:
- proteins
- Biology
- classifier
Model info
A finetuned Bert-Base-Uncased model for a multilabel classificaton task involving predicting protein functions based on their amino acid sequences. The model takes sequence data and protein class name as input and outputs probability scores. (How likely is it that this sequence belongs to this group)
Model Usage
The way you use this model for a demo is, you paste a protein sequence into the inference box and it outputs the relevant probabilities that certain GO terms are associated with that sequence. For example MMSTTHLLVFLLGVVTLTTPTFGTYESPNYGKPPTPVFKPPKVKPPPYEPKPPVYEPPKKEKPEPKPPVYAPPKKEKHGPKPTMYEPPKKEKPEPKPPVYTPPKKEVPKPKPPVYEPPKKEKPEPKPPIYTPPKKEKPEPKPPVYEPPKKEKPEPKPPVYTPPKKEKPEPKPPVYEPPKKPPMYEPKPPKPPVYTPPKKEKPEPKPPMYEPPKKPPMYEPKPPKPPVYTPPKKEKPEPKPPMYQPPNNPPIYEPKPPKPPVYAPPKEEKPKPKPPVYEPPAHEPPYGHYPGHPPLGKPQ