Drug discovery is a time-consuming and costly process, particularly given the high number of trials that ultimately “fail” or have negative outcomes. A high percentage of negative trial outcomes is to be expected in early-phase (I or II) trials, given they are mostly used as a proof-of-concept. However, the estimated 50% negative outcome rate in phase III trials represents a significant burden of cost in the drug development pipeline – and is a key target of genomics research.
Why do drugs fail?
To better understand the role of cancer genomics in the drug development pipeline, it is important to understand why so many drugs fail to reach market. Six key “drivers of failure” for clinical trials have been proposed:
1. Inadequate basic science
2. Suboptimal dose selection (inadequate dose finding studies or narrow therapeutic index)
3. Insufficient assessment of the current standard of care and disease area landscape
4. Flawed study design
5. Flawed data collection/analysis
6. Study operations problems
It is equally important to get things right at the research and discovery stage, and this is where genomics can play a hand in improving outcomes. Despite the extensive research and vetting for disease activity that goes into developing new molecular entities (NMEs) in drug development, it is estimated that around 90% of candidate NMEs in early-stage clinical trials fail to reach market. A key factor in a drug failing to reach approval is the cost of R&D. To maintain sustainability in the drug development process, pharma companies must increase the number of successful NMEs and reduce the number of failures. But this is no easy task.
Facilitating drug development with genomics
There are various ways that genomic information can help accelerate and improve drug development. Conceptual approaches in genetics and genomics help with target identification, prioritisation, and tractability, as well as predicting outcomes of pharmacological perturbations. Population genomics initiatives can also aid in target identification. Bulk and single-cell gene expression data is useful to understand the biological relevance of drug targets. Genome-wide CRISPR editing can screen for loss of function or activation of genes – a valuable tool for prioritising drug targets.
Genome sequencing and genotyping
Genome-wide association studies (GWAS) use high-density genotyping of common variants and linkage analysis. Exome sequencing captures the coding region of the human genome (about 1.5% of the entire genome). Whole genome sequencing achieves good coverage (around 85%) of the whole genome.
Exome sequencing and WGS are useful in identifying specific rare disease-associated variants that may be causal in cancer. Technical specifications of each technology may determine their success in translating variant discovery into actionable targets.
Table 1 outlines the strengths and limitations of these techniques with regards to drug discovery and development.
Transcriptomics
Transcriptional profiling of cells and tissues is a common technique in drug discovery, with high relevance to cancer drug and therapeutic development. Its use in supporting drug development includes mapping responses to compounds, interrogating tissues and cells for expression of target variants, and identifying causal variants of clinical phenotypes. It can also be used as a source of biomarkers to stratify patients for clinical trials.
Transcriptomics offers insights into the mechanisms of action and off-target effects in drugs. RNA-sequencing is not constrained by cell types or numbers, meaning accurate physiological models can be selected. This flexibility is derived from protocols ranging across low inputs, bulk or single-cell interrogation, and spatial transcriptomics.
Biomarkers can be identified with transcriptomics and are useful for cohort stratification and prediction of therapeutic outcomes. In cancer genomics-informed drug discovery, tissues and cells derived from biopsies are profiled by RNA-seq and the gene sample expression matrix is fed to supervised machine learning algorithms for classification and regression.
CRISPR-based technologies
CRISPR-based genome editing facilitates the creation of targeted genetic perturbations at scale and can screen for a phenotype of interest. RNA programmable genome-targeting by CRISPR/Cas-9 has been used to inhibit or activate transcription, edit nucleotides, and modify epigenetic states.
Screening for disease-relevant or drug mechanism-of-action targets are limited by suitability and scalability of available model systems. CRISPR screens have, nonetheless, driven target prioritisation for various disease models and clarified targets, enhancers, and resistance genes for existing drugs.
