Overview of Big Data in Biology
Treatment of target diseases
How does machine learning work in biology?
Applications of machine learning in biology and medicine?
References
References
Today, vast amounts of data are being generated in biological laboratories around the world. These can arise from gene sequence analysis, metabolome analysis, etc. Advances like these have allowed us to advance our understanding of the complexities of human biology and disease.
More recently, it has become necessary to use more advanced analytical techniques to be able to mine such data for greater profit. Machine learning is a next-generation technology and a subset of artificial intelligence (AI) currently used to navigate complex biological information while searching for specific patterns. This makes machine learning an ideal tool for targeted therapy in medicine.
Image credit: poptica/Shutterstock.com
Big data biology overview
The last decade has seen a dramatic increase in the number of large and highly complex datasets generated in biological experiments. These capture population-wide genetic variation in genes, proteins, metabolite abundances, microbiome composition, and among other variables. As Camacho et al. (2018) state, “We are living in an era of big data in biology and medicine, where data are collected at different layers of biological organization.”
Collaborators currently participating in the field of biological big data experiments typically generate petabyte-scale data (1 petabyte equals 100 million ((1015)) or, more precisely, 250 bytes) data. For example, The Cancer Genome Atlas (TCGA) has generated 2.5 petabytes of genomic, transcriptome, proteome, and epigenome data. This groundbreaking cancer genomics program sampled multiple omics measurements across 33 different cancer types.
Image credit: Jirsak/Shutterstock.com
Treatment of target diseases
Targeted disease treatment is a form of precision medicine. A growing area in this area is targeted therapy, cancer treatments that specifically target proteins that control cancer cells. This treatment involves the use of drugs or other substances that “target” and destroy or destroy carcinogenic cells. Meanwhile, healthy cells in close proximity to the pathological area remain intact. Common targeted therapies include monoclonal antibodies and small molecule drugs.
How does machine learning work in biology?
Machine learning is a branch of AI and computer science that uses algorithms to mimic the way humans learn. This includes the use of computer software that can be learned and adapted without the need for programming. Using algorithms and statistical data, machine learning can draw inferences from patterns in data. Practical examples include voice search technology and image recognition. In life science laboratories, machine learning has become an ideal tool for navigating the large biological datasets that are currently being generated in large numbers.
Machine learning-based AI can be used to detect new cancer targets. Applications consist of classification, clustering and neural networks. Two of the traditional machine learning-based algorithms are (1) decision trees and (2) deep learning. The decision tree algorithm works by selecting topological features of the cancer. Here, a supervised classification algorithm (which involves using training data that has already been labeled or classified) is employed. This means that specific biomarkers (such as genes or proteins) can be classified as primary targets. Such taxonomy-based applications now utilize genome-wide transcription profiles, protein expression profiles and/or mutational landscapes to classify tumor subtypes with high accuracy.
Deep learning algorithms use neural network capabilities (artificial networks that mimic the biological neural circuitry of the human brain) for cancer target identification and drug discovery. Many neural network models are currently being deployed in machine learning-based analytics. They benefit from a powerful ability to mine complex biological information via links or nodes (i.e., interconnected “neurons” modeled after the human brain).
The identification and annotation of genes within newly sequenced genomes provides a concrete example of machine learning in a biological context. Here, machine learning algorithms can learn about the genome and its key features, such as transcription start sites and specific genomic characteristics of genes such as GC content. This knowledge is used to generate models to find these key properties. The algorithm can apply what it learns from training data to entirely new genomes to make predictions about organization and functional capacity.
Artificial Intelligence in Medicine: Opportunities and Challenges | Navid Tusi Saidi | TEDxQUT
Applications of machine learning in biology and medicine?
Recent developments in cancer-related multi-omics technologies are crucial for the discovery of new anti-cancer targets and are fully coupled with AI biological analysis.
An application of machine learning that is becoming more widely used in biology involves genome annotation. Prediction of protein binding. Identification of major transcriptional drivers in cancer. Predicting metabolic functions and characterizing transcriptional regulatory networks in complex microbial communities (Camacho, other., 2018). In fact, any task in which patterns can be learned and applied to new datasets can be targeted for machine learning.
