Description Target Audience Prerequisites Outline

Course Description

This workshop is intended to provide an introduction to machine learning and its application to bioinformatics. This workshop is not intended for machine learning experts. Instead, it targets biologists or other life scientists who want to understand what machine learning is, what it can do and how it can be used for a variety of bioinformatic or medical informatics applications.

Course Objectives

Students will gain experience in:

Applications and Limitations of Machine Learning and Deep Learning
Decision Trees and Random Forests – how they work, how they are coded in Python and R, and how they can be used in bioinformatic applications (biomarker discovery and modeling)
Artificial Neural Networks (ANNs) – how they work, how data is encoded, how they are coded in Python and R, and how they can be used in bioinformatic applications (classification and secondary structure prediction)
Hidden Markov Models (HMMs) – how they work, how they are coded in Python and R and how they can be used in bioinformatics applications (gene finding)
Using Machine Learning tools (Decision Trees, ANNs and HMMs) on the Web (SciKit Learn and Keras/Colab)

Target Audience

Graduates, postgraduates, staff bioinformaticians and PIs working with or about to embark on using machine learning for bioinformatics applications.

Prerequisites

Familiar with the basics of Python and/or R.

You will also require your own laptop computer. Minimum requirements: 1024×768 screen resolution, 1.5GHz CPU, 2GB RAM, 10GB free disk space, recent versions of Windows, Mac OS X or Linux (Most computers purchased in the past 3-4 years likely meet these requirements).

This workshop requires participants to complete pre-workshop tasks and readings.

Course Outline

Module 1: Introduction to Machine Learning

What it is, what it isn’t. What it can be used for, what it shouldn’t be used for.
Examples of machine learning in bioinformatics
Brief introduction to machine learning methods including:
- 1) artificial neural networks
- 2) hidden markov models
- 3) decision trees and random forests
- 4) deep neural networks

Module 2: Decision Trees and Random Forests

Details of how a simple Decision Tree (DT) works and how a Random Forest (RF) works. An example of using DTs in classification.
Coding (Python and R) a simple DT to do classification
Example of RF in regression
Assessing model performance

Module 3: Artificial Neural Networks

Details of how a simple ANN works
The meaning of hidden layers
An example ANN simulation
Coding (Python and R) a simple ANN to do classification
Assessing model performance

Module 4: Artificial Neural Networks for Secondary Structure Prediction

Coding (Python and R) a simple ANN to do secondary structure prediction
Assessing model performance

Module 5: Gene Prediction with NNs

Review the structure of prokaryotic genes
Explain how prokaryotic gene prediction or identification can be done and how gene prediction is assessed
Review and explain the python code for different ANNs to predict prokaryotic genes
Use Colab to explore gene prediction code

Modules 6/7: Machine Learning with Keras and Scikit-learn

Introduce scikit-learn and Keras, compare to numpy
Show how the Iris classification problem with decision trees can be simplified using scikit-learn (sklearn)
Show how the ANN Iris classifier can be coded using Keras and sklearn
Use Colab to explore different code sets