Table 1 Comparison with other metal predictors

From: PinMyMetal: a hybrid learning system to accurately model transition metal binding sites in macromolecules

Predictor

Category

Input data

Method

Output data

Type and number of ligands

Provide metal ion location

Provide a structural model

Typical response time

Year of publication

PMM

III

Structure, Uniprot ID

Geometry, ML

PDB file, Structure

CHED ≥ 2

Yes

Yes

5–50 s

2023

Metal3D

II

Structure

CNN

Zinc ion location

N/A

Yes

No

3-60 min

2023

AlphaFill

II

Structure, Uniprot ID

Structure homology

PDB file, Structure

N/A

Yes

Yes

5–50 s

2023

ZincBindDB

I

Structure, Sequence

ML

Predicted sites

CHED ≥ 2

No

No

3-10 min

2021

znMachine

I

Sequence

ML

Predicted sites

CHED ≥ 3

No

No

Unavailable

2019

GRE4Zn

III

Structure

Geometric restriction

PDB file

CHED ≥ 3

Yes

No

5–30 s

2014

TEMSP

III

Structure

ML

PDB file

CHED ≥ 3

Yes

No

Unavailable

2011

CHED

I

Structure

ML

Predicted sites

CHED

No

Yes

Unavailable

2007

  1. Categories (I) binding site predictors for metal binding residues; (II) binding position predictors for metal ion coordinates; (III) predictors that identify both residues and coordinates.
  2. Metal3D’s runtime can be accelerated using GPU processing.