Skip to main content

ProtNLM for Function Prediction

ProtNLM predicts the function of a protein based on its amino acid sequence. It is a T5-based model developed by Google which returns a number of suggested descriptive names, rank ordered by their score (0.0, 1.0) where larger is better. These predictions help scientists explore the potential functions of new proteins and evaluate computer-designed proteins, with sequence comparison to known proteins as a common first step.

How to Use ProtNLM

Predict the function of ‘DIHICGICKQQFNNLDAFVAHKQSGSQ’

NameScore
C2H2-type domain-containing protein0.54
General transcription factor II-I repeat ...0.02
Zinc finger protein0.019

Inputs

Protein Sequence

A string representing the single-letter amino acid sequence, consisting of the common 20 amino acids. This input is intended for single-chain (monomeric) proteins only.

Outputs

Text Descriptions

Short text descriptions of the protein, ranked by confidence scores (0.0 to 1.0).

Confidence Score

Confidence level for each description (0.0 to 1.0). Higher scores are better.

Examples

Predict the function of M4332

NameScore
Tyrosine-protein kinase0.592
2.7.10.20.249
Non-specific protein-tyrosine kinase0.019

Load Q9UPY3 and predict its function

NameScore
Endoribonuclease Dicer0.466
3.1.26.30.409
Endoribonuclease Dcr-10.033