#
# EASE-MM: sequence-based prediction of mutation-induced stability changes with feature-based multiple models
#
# This file describes the output format for the EASE-MM web-server.
#

###########
# EASE-MM #
###########

EASE-MM stands for Evolutionary, Structural, and Amino acid Encodings with Multiple Models.
It is a machine learning method to predict the real value of the stability change, denoted as ∆∆Gu (also referred to as ddG).

Please note that if you are using "EASE-MM Genome", only missense variants are supported. All other types of genetic variants are ignored during prediction.

###########################
# ∆∆Gu - stability change #
###########################

EASE-MM uses the same convention as the ProTherm database, which defines the stability change as the difference in the Gibbs free energies of unfolding between the mutated and wild-type proteins:
∆∆Gu = ∆Gu(mutated) - ∆Gu(wild-type).
Therefore, ∆∆Gu < 0 means that a mutation decreases the protein stability (a destabilising mutation).

###################
# Stability class #
###################

A very simple classification of predicted stability changes is adopted here based on the predicted value of ∆∆Gu:
* ∆∆Gu in (-inf, -1]  --> destabilising
* ∆∆Gu in (-1, -0.5]  --> likely destabilising
* ∆∆Gu in (-0.5, 0.5) --> neutral
* ∆∆Gu in [0.5, 1]    --> likely stabilising
* ∆∆Gu in [1, +inf)   --> stabilising

####################################################################################
# SS and rASA - predicted secondary structure and relative accessible surface area #
####################################################################################

The last two columns in the output format are the predicted secondary structure (SS) and relative accessible surface area (rASA) of the mutation site.
SS [helix (H), extended/sheet (E), or coil (C)] and rASA were calculated with SPIDER. Disorder probability was estimated with SPINE-D.
Evolutionary sequence conservation was extracted from the PSSM output of PSI-BLAST.