Discovery research is the earliest stage of the drug development process, where scientists work to understand the biological basis of a disease and find molecules that could potentially treat it. It spans everything from identifying a protein or gene involved in a disease to optimizing a chemical compound that can act on that target. This phase typically takes five to six years and sits before any testing in humans begins.
Where Discovery Fits in Drug Development
Drug development is generally broken into five stages: basic research and pre-discovery, drug discovery, preclinical development, clinical trials, and regulatory approval. Discovery research covers the first two of these. During basic research, scientists study how a disease works at the molecular level and propose biological targets, usually specific proteins, that play a role in the disease. During the discovery stage itself, they search for molecules that can interact with those targets to cure, slow, or ease symptoms.
Once a promising molecule is found, it moves into preclinical development, a two-to-three-year phase focused on safety testing in animal models and early formulation work (figuring out whether the drug will be a pill, injection, or something else). Only after that does a drug candidate enter clinical trials in humans. Discovery research is the foundation the entire pipeline rests on, and its quality largely determines whether later stages succeed or fail.
What Happens During Discovery Research
The process follows a loose sequence: target identification, target validation, assay development, screening, hit identification, lead optimization, and candidate selection. In practice, these steps overlap and circle back on each other constantly.
Target identification means pinpointing a biological molecule, often a protein, that contributes to a disease. Scientists use genomic data, studies of diseased tissue, and computational modeling to find these targets. Target validation then confirms that interfering with that molecule actually affects the disease in a meaningful way, usually through cell-based experiments or animal models.
Once a validated target exists, researchers build assays (standardized laboratory tests) designed to measure whether a compound interacts with that target. These assays feed into high-throughput screening, where automated systems test enormous libraries of chemical compounds against the target. Modern high-throughput screening platforms can evaluate hundreds of thousands of compounds in a single day, replacing the older trial-and-error approach.
Compounds that show activity in screening are called “hits.” Most hits aren’t ready to become drugs. They may bind weakly to the target, break down too quickly in the body, or cause unwanted effects. Through rounds of medicinal chemistry, scientists modify these hits to improve their properties, eventually producing “leads.” Lead optimization refines these further until a single candidate molecule is selected to move forward into preclinical testing.
Why Most Discovery Projects Fail
Discovery research has the lowest success rate of any stage in drug development. Only about 32% of projects that begin at the preclinical discovery stage successfully produce a candidate that advances further. For context, later stages have considerably better odds: roughly 75% of drugs that enter Phase I clinical trials survive that phase, and about 59% make it through Phase III.
The high failure rate has major financial consequences. When looking only at direct costs, the nonclinical stage accounts for about 7% of the average drug’s total development expense. But when you factor in the cost of all the failed projects that never produced a viable drug, that share jumps to 27%. Include the cost of capital tied up during those years and it rises to roughly 40% of the total. One analysis of U.S. drug development from 2000 to 2018 estimated that the average cost to develop a single new drug was about $173 million in direct spending, but climbed to $879 million once failures and capital costs were included.
Who Does Discovery Research
Both academic institutions and pharmaceutical companies conduct discovery research, but they play different roles. Historically, large pharma companies ran the entire process in-house. That model has shifted. Companies increasingly partner with universities and academic drug discovery centers in early-stage research to share risk and access emerging technologies. NIH funding contributed to identifying or understanding the mechanism behind 354 of 356 new drugs approved in the U.S. between 2010 and 2019, a remarkable 99.4%.
Academic labs bring strengths that industry often lacks. They can pursue high-risk ideas without the immediate pressure to generate profit, and they tend to focus on diseases with limited commercial appeal that pharmaceutical companies avoid. One illustrative case: academic researchers revived a promising cancer compound that had been abandoned by industry due to side effects, redesigned it to reduce toxicity, and advanced it into early clinical trials. That kind of long-shot project is rare in a corporate setting where resources get pulled from unpromising programs quickly.
The collaboration isn’t always smooth. Universities prioritize publishing findings, while companies prioritize speed and commercialization. Intellectual property disputes, misaligned timelines, and cultural differences between the two worlds can slow things down. Academic centers also face funding instability, relying on a patchwork of government grants, philanthropic donations, and industry partnerships rather than the large, predictable R&D budgets that pharmaceutical companies plan years in advance.
How Technology Is Changing Discovery
Two tools have reshaped discovery research over the past two decades: high-throughput screening and computational methods. High-throughput screening replaced manual, one-compound-at-a-time testing with automated platforms that can scan massive chemical libraries against a target. Compounds can be tested virtually, using computer-simulated molecular structures, or physically through cell-based and biochemical assays. The “hits” from these screens become the raw material for further chemistry work.
Machine learning has added another layer. Algorithms trained on molecular data can predict which compounds are likely to interact with a target, narrowing the field before any physical screening begins. This approach has shown particular promise in antibiotic research, where traditional discovery methods have stagnated. Whole-genome screening techniques also allow researchers to systematically test every gene’s role in a disease, identifying new targets that might not have been obvious from studying the disease alone.
Together, these tools have compressed some parts of the discovery timeline and opened up target classes that were previously considered too difficult. But the fundamental challenge remains: biology is complex, and a molecule that looks promising in a lab assay often behaves differently in a living system. That gap between the dish and the body is the central reason discovery research stays risky, expensive, and slow.