Key points
- Most modern drug discovery starts with a biological target: a protein, gene, or pathway believed to drive a disease.
- Researchers screen for 'hits' that act on the target, then refine them over many cycles into a 'lead' and finally a development candidate.
- Before human testing, a candidate goes through preclinical pharmacology and toxicology studies, and the sponsor files an Investigational New Drug (IND) application with the FDA.
- Targets backed by human genetic evidence are substantially more likely to succeed in clinical development, which is one reason genomics now sits at the center of discovery.
Drug discovery is the work that happens before a medicine is ever given to a person. It answers two questions: what should we try to change in the body, and what molecule can change it safely? Most candidates are eliminated during this stage, long before a clinical trial is registered.
Step 1: Choosing a target
A target is the biological structure a drug is meant to act on. It is usually a protein, such as an enzyme, a receptor, or an ion channel, but it can also be a gene, an RNA molecule, or a whole signaling pathway. Researchers look for a target whose activity is linked to a disease and whose modulation should help patients.
Picking the right target is arguably the most consequential decision in the whole process. A molecule can hit its target perfectly and still fail in patients if the target turns out not to matter for the disease. That is why researchers try to validate a target before investing heavily, using evidence such as:
- Human genetics. People who carry natural variants that disable or enhance the target have different disease risk.
- Disease biology. The target is overactive, underactive, or misplaced in diseased tissue.
- Experimental models. Blocking or activating the target in cells or animals changes the disease.
- Clinical precedent. Other drugs acting on the same pathway have already shown an effect.
Why genetics matters so much
Genetic evidence is one of the few validation signals with a measurable track record. A 2015 analysis in Nature Genetics found that drug mechanisms supported by human genetic associations were considerably more likely to reach approval. A larger follow-up published in Nature in 2024 estimated that drug programs with genetic support were about 2.6 times more likely to succeed from clinical development to approval than programs without it.
This is a large part of why genomics, the study of an organism's complete set of genes, has moved from a research specialty to the center of drug discovery.
Step 2: Finding a hit
Once a target is chosen, researchers look for a starting molecule that interacts with it. A molecule that shows the desired activity in an initial test is called a hit. Common ways to find hits include:
| Approach | How it works |
|---|---|
| High-throughput screening | Automated testing of large chemical libraries, often hundreds of thousands to millions of compounds, against the target. |
| Fragment-based discovery | Screening very small chemical fragments that bind weakly, then growing or linking them into stronger binders. |
| Structure-based design | Using the three-dimensional structure of the target to design molecules that fit its binding site. |
| Biologic discovery | Generating antibodies or other proteins against the target, for example by immunization or display technologies. |
| Computational screening | Using software models to predict which molecules are likely to bind, before testing the best candidates in the lab. |
Step 3: From hit to lead to candidate
A hit is rarely good enough to become a medicine. In the hit-to-lead and lead optimization phases, chemists and biologists make and test many variations of the molecule. Each cycle aims to improve several properties at once:
- Potency: how strongly the molecule acts on the target.
- Selectivity: how little it acts on other targets, which reduces side effects.
- Pharmacokinetics: how the body absorbs, distributes, metabolizes, and excretes it (often shortened to ADME).
- Safety: early signals of toxicity in cells and animals.
- Manufacturability: whether it can be produced consistently at scale.
Improving one property often worsens another, so lead optimization is a balancing act. The result is a single development candidate, sometimes with one or two backups.
Step 4: Preclinical studies
Before a candidate can be tested in people, the sponsor must generate enough laboratory and animal data to show it is reasonably safe to proceed. Preclinical work typically includes:
- Pharmacology studies showing the drug has the intended biological effect.
- Toxicology studies, usually in more than one animal species, to identify the organs that might be harmed and the doses at which harm occurs.
- Manufacturing and quality data describing how the drug is made and controlled.
In the United States, the toxicology studies that support human testing are generally run under Good Laboratory Practice rules (21 CFR Part 58), which set standards for how studies are conducted and documented.
Step 5: The Investigational New Drug application
To begin clinical trials in the United States, the sponsor submits an Investigational New Drug (IND) application to the FDA. The IND contains the preclinical data, manufacturing information, and the plan for the first human study. Under the IND regulations (21 CFR Part 312), the sponsor may generally begin the trial 30 days after the FDA receives the application, unless the FDA places it on clinical hold.
The IND marks the handoff from discovery to clinical development, which is explained in What Are the Phases of Clinical Trials?.
Different kinds of drugs, different paths
The steps above describe a classic small-molecule program, but the same logic applies to other modalities with different tools:
- Small molecules: chemically synthesized compounds, often taken as pills.
- Monoclonal antibodies and other proteins: large biologic molecules produced in living cells.
- Oligonucleotides: short strands of DNA or RNA that modulate gene expression.
- Gene therapies: treatments that deliver, edit, or silence genetic material.
- Cell therapies: living cells, such as engineered immune cells, given as treatment.
The modality affects how a drug is discovered, how it is manufactured, which FDA center reviews it, and, as later guides explain, its historical odds of success.
What this means when you read about a new drug
When a company announces a "first-in-class" candidate, the most useful questions are often about discovery decisions made years earlier:
- How strong is the evidence that the target drives the disease in humans?
- Is there human genetic support for the mechanism?
- Have other drugs acting on the same target succeeded or failed?
- What did the preclinical safety studies show?
Spectral Genomics profiles of experimental medical drugs are designed to answer these questions alongside the clinical data.
Sources
- U.S. Food and Drug Administration. The Drug Development Process.Regulator
- Hughes JP, Rees S, Kalindjian SB, Philpott KL. Principles of early drug discovery. British Journal of Pharmacology. 2011;162(6):1239–1249.Peer-reviewed
- Nelson MR, et al. The support of human genetic evidence for approved drug indications. Nature Genetics. 2015;47:856–860.Peer-reviewed
- Minikel EV, Painter JL, Dong CC, Nelson MR. Refining the impact of genetic evidence on clinical success. Nature. 2024;629:624–629.Peer-reviewed
- 21 CFR Part 58: Good Laboratory Practice for Nonclinical Laboratory Studies.Statute / regulation
- 21 CFR Part 312: Investigational New Drug Application.Statute / regulation
This article is for education. It is not medical advice and not investment advice. Figures are drawn from the sources above and describe historical averages, not the prospects of any particular drug. See our editorial standards and disclaimer.