Pharmacogenomics improves research into drug efficiency, pertinent dosage, and genetic risks for conflicting drug reactions. Recent developments in medical sciences and pharmacogenomics concentrate on the effective, faster, and economical scope of drug delivery. Bioinformatics, DNA/RNA sequence analyses, Pharmacology and healthcare, and chemo-informatics for drug design are the impressive deep learning-enabled pharmacogenomics research areas. Pharmacogenomics identifies genetic variants linked with drug effects in populations, cohorts, and individual patients.
Pharmacogenomics delivers promise for applications such as medicament optimization for patients based on genotype in diagnostic testing, drug discovery, and development. Recently, machine learning and deep learning techniques have been utilized for pharmacogenomics. Deep learning models have gained outstanding performance in several pharmacogenomics applicative tasks: predicting DNA accessibility within noncoding regions, possible transcription factor and RNA binding sites and gene expression from histone alterations, analysis of noncoding regulatory, pharmacogenomic patient stratification, personalized treatment outcome prediction, and medication optimization.
Some trending applications of deep learning in pharmacogenomics are cell phenotypes prediction in transcriptomics data, drug response in cancer, seizure-persuading side effects of preclinical drugs, patient survival detection from multi-omics data, predicting drug-induced liver injury and genomic variants classification into adverse drug reactions.
Discovery in research & development, clinical decision support, pharmacogenomic variation, clinical trial participants, drug response, and drug-gene interaction pharmacogenomics are the trending topics in deep learning-based pharmacogenomics. Some of the future scopes of deep learning-based pharmacogenomics are data-driven surrogate phenotypes, prediction of new variants from the noncoding regulatory genome, cohorts stratified by deep learning-predicted similar pharmacogenomic risk, prediction of dosage using deep learning, joint learning from pharmaco-omics data. Data requirements, overfitting, and interpretability are some notable challenges in deep learning-based pharmacogenomics.
Various surveys and reviews in deep learning-enabled pharmacogenomics cover methodological advantages, opportunities, challenges, limitations, futuristic applications, benchmark datasets, research directions, and emerging strategies.