Deep Learning for Life Sciences with Python Notebooks for Examples and Exercises

Artificial intelligence is already ubiquitous in the life sciences, from cancer diagnosis to medical image analysis, from precision agriculture to wildlife monitoring. It is therefore essential for any scientist, especially life scientists, to have a basic understanding of deep learning, the statistical engine behind AI.

This book explains the theory behind neural networks and their internal workings in clear terms and with numerous examples. The authors cover the building blocks of neural networks, the mathematical theory, different types of network architectures, the problem of overfitting, and the strategies to avoid it. The most common data types encountered in biological problems are discussed, with suggestions on how to apply deep learning to different cases. Success and failure stories are presented through interviews with leading experts in the field.

The book is accompanied by several Python notebooks with practical examples and clearly commented code.

Introduction.- Part 0.- From statistics to statistical learning.- Part 1.- Neural networks and deep learning decoded.- Part 2.- Making it work.- Part 3.- Deep learning models for biological research.- Part 4.- Success and failure cases of deep learning applications in biology.- Appendixes.

Filippo Biscarini is a Senior Scientist at the CNR (Consiglio Nazionale delle Ricerche) in Milan, Italy. He holds a Ph.D. in Animal Breeding and Genomics from Wageningen University (The Netherlands). He has worked in several countries (Italy, Ireland, Canada, The Netherlands, Germany, Spain, UK, Belgium) for the industry, academic institutions and international organizations (European Union, United Nations). His research activities include bioinformatics and biostatistics applications in plant and animal sciences and in human medicine. He has coordinated several national and international research projects. Current research activities include neural network models for the analysis of genomic and phenomic data in agriculture, the analysis of the composition and role of microbiomes across life domains, and the development of predictive models for precision medicine.

Nelson Nazzicari is a Senior Scientist at CREA, Italy's largest agricultural research institute. He holds a Ph.D. in Informatics and Electrical Engineering from the University of Pavia (Italy), spent two years at George Mason University (VA, USA), taught programming at the University of Mantua and then transitioned from pure informatics to bioinformatics. Nelson is affiliated with the Faculty of Agriculture at the University of Zagreb (Croatia) and collaborates with the Netherlands Plant Eco-phenotyping Centre (Wageningen, NL). His current research focuses on genotype-phenotype modeling and the use of deep learning in precision agriculture.