[Cbe-grads] Information about CBE 5734 (Molecular Informatics) SP 2025

[Rathman, Jim]

Hi everyone,
As you plan your schedule for Spring 2025, I encourage you to consider my Molecular Informatics technical elective, described below. Please note that prior computer programming experience is expected. We use Python in this class so prior experience with Python is recommended.

Professor Jim Rathman
Chemical & Biomolecular Engineering

----------------------------------------------------------------------
CBE 5734: Molecular Informatics
Credits: 3
Term: Spring 2025
When: TuTh 9:35 - 10:55 am
Where: CBEC 110/120
Mode: Hybrid – mixture of in-person and online (Zoom) class sessions
Textbook: none

Molecular informatics is the combination of science, engineering, information science, and computers to develop in silico (computational) approaches for understanding and predicting properties and effects of molecules in chemical and biological systems. Bioinformatics is molecular informatics applied to proteins, DNA and RNA, whereas chemoinformatics is molecular informatics applied to chemical components of drugs, foods, pharmaceuticals, agrochemicals, etc. Although we will briefly discuss bioinformatics, this course will focus on chemoinformatics. Lacking experimental evidence, how can we determine whether a given compound might have chemical toxicity effects that could negatively impact human health or raise ecotoxicity concerns? Chemoinformatics helps us answer such questions by developing in silico models to predict specific outcomes, where the models are trained and validated using data available for molecules similar to the compound of interest.

Learning how to write computer programs is a key skill in molecular informatics. We will use the Python programming language. You are expected to have prior programming experience such as CBE 2345 or CBE 5790 or an introductory programming course in C+ or MATLAB or Python. Additional resources, such as tutorials and special office hour sessions, will be available early in the semester if you need help learning or reviewing Python programming fundamentals.

CBE 5734 – Molecular Informatics, SP 2025
Topics
Descriptions
Chemical structure representation
How chemical structures can be represented in computable form; 2D and 3D representations; stereochemistry.

  *   SMILES, InChI codes and keys
  *   MOL and SD files

Chemical descriptors
Descriptors extracted from chemical structures for use in characterization and modeling

  *   structural fingerprints
  *   computed molecular properties (e.g., hydrogen bond donors and acceptors)
  *   computed physicochemical properties (e.g., logP)

Chemical similarity
Methods for quantifying the similarity between two compounds taking into account

  *   structure
  *   properties
  *   biological activity

Chemoinformatics databases
Structure searching strategies for

  *   exact match
  *   similarity search
  *   substructure search (SMARTS)

Computational methods to predict chemical properties and activity
Profiling

  *   Structural rules and alerts

Predictive modeling of property or activity endpoints

  *   (Q)SPR – (Quantitative) structure-property relationships
  *   (Q)SAR – (Quantitative) structure-activity relationships
  *   Rule-based approaches

Machine learning

  *   Key concepts: cross-validation, external validation, feature scaling, regularization, regularized linear models, hyperparameters and hyperparameter tuning, latent variable approaches (PLS, PCA, …)
  *   Multivariate regression
  *   k-nearest neighbors regression and classification
  *   Logistic regression
  *   Support vector machines for classification and regression
  *   Decision trees





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osu.edu/pipermail/cbe-grads/attachments/20241014/806ff414/attachment.html>