The aim of this module is to provide students with skills and knowledge to apply computational techniques to analyse and model chemical data for molecular and materials design and discovery. The module will be delivered using a combination of lectures and workshops to give students the experience of applying their skills to a range of chemistry contexts.

(LO1) Explain the difference between statistical models and physics-based models, and provide examples of these approaches pertinent to different areas of chemical research including that of their research project

(LO2) Ability to describe the most common methodologies used in chemical research to perform high-throughput virtual screening and how they are validated.

(LO3) Extract relevant information from databases of molecular or materials properties

(LO4) Use computer coding skills to develop workflows that accelerate the analysis of experimental or computational chemical data in reproducible and transparent way

(LO5) Deploy the most common ML methods to derive predictive models for chemical data analysis and design, assess their quality and propose strategies to generate improved models based on further data collection or ML methodology

(S1) IT (coding, graphical tools), data analysis skills

(S2) Critical analysis

(S3) Presentation skills

Lectures: 8 hr
Concepts will be delivered in 1 hr lectures to provide the underlying knowledge to tackle the chemical data scenarios developed in the practical workshops.

Workshops: 8 x 3 hr
Building on students’ coding skills, including those acquired in Semester 1 module COMP517, computational workshops will be used to give students practical experience in applying the concepts covered in lectures to a variety of chemical data contexts. The workshops will include elements of data retrieval and curation, the implementation of methods for data analysis and their critical evaluation.

Chemical descriptors and data sets
- Representations of molecules and materials as chemical data
- Programmatically accessing and exploring repositories of molecular data and their properties

Introduction to Data Science methods for chemical data
- General introduction to the main machine learning (ML) and artificial intelligence (AI) approaches applied to chemical data.
- ML and AI approaches for chemical data challenges, including areas such as data reduction, literature mining, optimisation, similarity analysis, classification and property prediction; specific examples of applications in chemistry in experimental and computational contexts.
- Considerations for method/model selection, including data availability, model complexity, interpretability. Critically assessing model performance and outputs, including against chemical and/or physical feasibility.
- Ethical considerations of AI in chemistry.

Elements of Computatio nal Chemistry for Virtual Discovery
- Evaluation of system energy through empirical potential
- Outline of Density Functional theory
- Simulation methods
- High-throughput computational approach (virtual discovery, crystal structure prediction)

Data Science and ML approaches to Chemical Materials Discovery
- Examples of recent applications for different areas of chemistry and materials.