MADRIGAL

The Madrigal model was developed by Staller Lab as an activation domain predictor. Given a protein sequence (or sequences) in a FASTA format or a plain text, it outputs the graph showing the likelihood of different regions to be part of an activation domain. If a sub-sequence of amino acids goes above a certain 'threshold' of likeliness, then it can be predicted to be an activation domain.

We recommend a threshold of 0.7 for results that are accurate and reduce the number of both false positives and negatives in inference data.

Please input your data as FASTA or plain text. An example is shown below.

Query: Enter as a string or FASTA

Or Upload a FASTA file

Optional: Set a threshold between 0-1.1 for valid sequence

Provided by Staller Lab 2026