Ligand-Protein Binding Affinity Prediction Using Machine Learning Scoring Functions - PhD Thesis Index

Aug 2, 2026

Upload by Eliana

This thesis investigates ligand-protein binding affinity prediction through machine learning scoring functions, building a research workflow from model design to evaluation. It covers core machine learning concepts, artificial neural network architectures, training methods including gradient descent and backpropagation, and strategies to address overfitting such as stopping criteria. A dedicated database is constructed using proteins, ligands, complexes, structures, affinity and docking scores, then processed with normalization and feature handling. The state-of-the-art is reviewed, and an MLP scoring function is developed, trained, and benchmarked via horizontal, vertical, and per-target test protocols with performance comparisons and related factors.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages136

Reading Time5 h 43 min

AccessFree to download

Introduction

Machine Learning

Artificial intelligence and machine learning

Artificial neural network description

Neural network training method

Neural network training strengthening: transfer learning

Neural network performance descriptors

Database

Database description

Experimental data

Synthetic data

Data distribution

Database creation result

State of the art

Machine learning scoring function

Database

Test types

Performance doping factor

MLP Scoring function

Ligand-protein complex descriptive model

MLP scoring function neural network and training protocol choice

Test set choice

MLP scoring function performance

Horizontal test

Vertical test

Per-target vertical test

Performance comparison

UNIVERSITA’ DEGLI STUDI DI CAMERINO School of Advanced Studies DOCTORATE COURSE IN “Physics” XXXIV Cycle TITLE OF THE THESIS: “Ligand-Protein Binding Affinity Prediction Using Machine Learning Scoring Functions” PhD Student Supervisor Ing. Francesco Pellicani Prof. Sebastiano Pilati Co-supervisor Prof. Diego Dal Ben Index Introduction ...................................................................................................................... 5 1 Machine Learning ....................................................................................................... 10 1.1 Artificial intelligence and machine learning ........................................................... 10 1.1.1 Artificial Intelligence ....................................................................................... 10 1.1.2 Machine learning ............................................................................................ 12 1.2 Artificial neural network description ...................................................................... 16 1.3 Neural network training method............................................................................22 1.3.1 Gradient descent ............................................................................................24 1.3.2 Gradients computation in artificial neural networks: backpropagation algorithm................................................................................................................................26 1.3.3 Overfitting and underfitting .............................................................................28 1.3.4 Stopping criterion ...........................................................................................32 1.4 Neural network training strengthening: transfer learning ......................................32 1.5 Neural network performance descriptors..............................................................33 2 Database .................................................................................................................... 36 2.1 Database description............................................................................................36 2.1.1 Proteins ..........................................................................................................38 2.1.2 Ligand ............................................................................................................43 2.1.3 Ligand-protein complex ..................................................................................44 2.1.4 Ligand-protein complex structure ...................................................................44 2.1.5 Ligand-protein complex affinity and docking score.........................................46 2.2 Experimental data ................................................................................................48 2.2.1 Experimental data preparation .......................................................................49 2.3 Synthetic data.......................................................................................................50 2.3.1 Synthetic data preparation .............................................................................51 2.4 Data distribution ...................................................................................................53 2.5 Database creation result ......................................................................................60 3 State of the art ............................................................................................................ 63 3.1 Machine learning scoring function ........................................................................63 3.2 Database ..............................................................................................................71 3.3 Test types ..............................................................

FAQs

What problem does the thesis address?

The thesis targets predicting ligand-protein binding affinity using machine learning scoring functions, and evaluates performance using multiple test types.

Which machine learning and neural network training concepts are covered?

It explains artificial intelligence and machine learning, neural network structure, training via gradient descent and backpropagation, overfitting/underfitting, stopping criteria, and transfer learning.

How is the evaluation of the MLP scoring function organized?

Performance is assessed through horizontal, vertical, and per-target vertical tests, followed by performance comparisons.

Aug 2, 2026

Upload by Eliana