Biotech science connects a disease process to a treatment designed to change it. For investors, the useful questions concern what the treatment acts on, how it reaches that target and whether the resulting biological change can improve patients’ health. A promising mechanism can support further research. It cannot establish that a medicine will work at a tolerable dose, remain effective or reach patients through a workable production process. This guide introduces the main drug technologies and the evidence needed to assess them, providing the scientific foundation for biotech investment research. The aim is to distinguish a plausible scientific idea from a development candidate whose remaining risks can be examined.
Separate the target, mechanism and treatment format
A drug target is the biological feature a treatment acts on, often a protein or RNA sequence. The mechanism of action describes what happens when the treatment acts on it. The treatment format, also called its modality, is the tool used to produce that effect: a small molecule, antibody or another technology. These distinctions matter because evidence supporting one part of the proposal does not establish the others. Researchers might connect a protein to a disease without having a suitable drug against it. The NIDDK framework for target identification places evidence linking the target to disease progression or symptoms near the beginning of therapeutic discovery.
Consider a hypothetical inflammatory disease in which an overactive protein sends harmful signals. An antibody might block the protein outside cells; an RNA medicine might reduce its production. Both approaches address related biology, yet they create different delivery, dosing and safety questions. Research should therefore separate evidence about the target from evidence about each candidate. Ask whether changing the target improves disease biology, whether the treatment produces that change and whether the change benefits patients. The guide to drug targets and mechanisms of action examines that reasoning in greater detail.
How the main drug technologies compare
Each technology changes what researchers can attempt and what investors should investigate. None earns a higher scientific rating simply because it is newer.
| Technology | Basic approach | Question to investigate |
|---|---|---|
| Small molecules | Chemical compounds that alter biological activity | Can sufficient drug reach the target without unacceptable toxicity? |
| Antibodies | Proteins designed to bind selected targets | Does binding produce the intended therapeutic effect? |
| Antibody–drug conjugates | Antibodies linked to drug payloads | Does payload delivery provide a useful benefit–risk balance? |
| Cell therapies | Living cells used as treatment | Do the cells function consistently after administration? |
| Gene transfer therapies | Delivery of genetic instructions into cells | Where are those instructions delivered, and how long does their effect last? |
| Gene editing | Tools designed to alter DNA sequences | Are the intended changes sufficient, with acceptable unintended effects? |
| RNA medicines | Modification of RNA activity or delivery of protein instructions | Can the medicine reach the required cells and produce useful activity? |
Small molecules and biologics
Small molecules are comparatively compact chemical compounds. Many can reach targets inside cells, making them useful tools for changing enzyme activity or cellular signalling. Their apparent simplicity should not be confused with easy development. A candidate still needs a workable balance between absorption, distribution, activity and clearance. Oral dosing is possible for many small molecules, but it is not automatic. For an investor comparing candidates, the relevant question is whether the compound’s properties suit the target and disease. A daily tablet may offer convenience, yet that advantage means little if the dose needed for efficacy also causes unacceptable adverse effects.
Biologics form a much broader category than antibodies alone. They include recombinant therapeutic proteins, certain vaccines, cellular products and gene therapies. Their composition and production methods create different analytical and handling demands from those of conventional chemically synthesised medicines. The FDA overview of biological products identifies their complexity, sensitivity to heat and susceptibility to microbial contamination. Avoid treating “biologic” as a single development profile. An antibody and a patient’s engineered cells present very different problems. The comparison of small molecules and biologics examines how these differences affect development and investment analysis.
Antibodies and antibody–drug conjugates
Therapeutic antibodies are proteins designed to recognise selected targets. Binding can block a biological signal, help immune cells recognise diseased cells or support another therapeutic action. Binding alone, however, is not proof of useful treatment activity. In cancer, antibodies can act against targets on cell surfaces, whereas small molecules can address targets inside cells. The NCI overview of targeted cancer therapies also documents resistance and serious adverse effects. “Targeted” describes the intended action; it does not guarantee that healthy tissue will escape harm. Investors should examine what the antibody does after binding and how that action connects to patient benefit.
An antibody–drug conjugate, or ADC, links an antibody to a drug payload through a chemical linker. Its investment case depends on the combined product, rather than the antibody’s binding properties alone. Research questions include whether the target is accessible, how the linker behaves and where the payload becomes active. Two candidates aimed at the same target can therefore deserve different assessments. A strong target rationale does not make their performance interchangeable. The guide to antibody therapies and antibody–drug conjugates examines those distinctions without treating the technology label as evidence of superiority.
Cell therapies: assess the cells and the treatment process
Cell therapies use living cells as the treatment. CAR T cell therapy illustrates how this differs from administering a conventional drug. In a patient-derived approach, T cells are collected, engineered to recognise a chosen target, expanded and returned to the patient. Their subsequent behaviour inside the body forms part of the treatment’s activity. The NCI account of CAR T cell engineering describes this process and risks including cytokine release syndrome and neurological complications. These risks make the surrounding treatment process relevant to scientific assessment. Investors need to examine how the cells perform and what supporting care patients require.
The starting material also matters. Autologous products use a patient’s own cells; allogeneic approaches use cells from another donor. These categories prompt different questions about consistency, immune compatibility and production planning. Do not assume that a donor-derived product automatically solves every constraint associated with patient-derived manufacturing. Ask which problems the design addresses and which remain unresolved. Useful evidence includes the proportion of attempted treatments that reach infusion, the consistency of the administered product and the duration of patient follow-up. The dedicated guide to cell therapies develops these scientific and operational questions.
Gene therapies and gene editing
Gene transfer introduces genetic instructions intended to restore or compensate for a biological function. Delivery can occur directly inside the body, called an in vivo approach, or through cells modified outside the body and then returned, an ex vivo approach. A carrier, or vector, helps deliver the material. Its behaviour belongs in the scientific assessment alongside the genetic instructions themselves. The National Library of Medicine explanation of gene therapy delivery distinguishes these approaches and describes risks associated with viral vectors. An attractive genetic rationale still leaves practical questions about which cells receive the treatment and what happens afterwards.
For investment research, separate the planned administration schedule from the demonstrated duration of benefit. A treatment intended for one administration has not, on that basis alone, established lifelong efficacy. Look for follow-up showing whether its biological effect and patient outcomes persist. Ask how researchers would detect declining activity, what evidence exists about later adverse effects and whether further treatment would be feasible. These are questions to test against the candidate’s data, not assumptions to fill with optimistic language. The guide to gene therapies examines delivery and durability in more detail.
Editing DNA does not always mean repairing the original mutation
Gene editing changes DNA at selected locations. Its therapeutic purpose can involve disrupting a sequence rather than directly correcting the disease-causing mutation. Vertex’s Casgevy provides a concrete example. On December 8, 2023, it became the first FDA-approved treatment using CRISPR/Cas9 genome editing. The treatment modifies patients’ blood-forming stem cells outside the body so that, after transplantation, they increase fetal haemoglobin production. The FDA announcement of the Casgevy approval also describes the chemotherapy conditioning required before infusion. The editing tool is one component of a demanding treatment process, not the whole product experience.
When assessing an editing candidate, distinguish the intended DNA change from the clinical outcome it is meant to produce. Examine how the company measures editing, which cells it samples and how it investigates unintended changes. A high editing percentage needs biological context. The guide to gene editing covers precision, delivery and uncertainty beyond the initial measurement.
RNA therapeutics act through different mechanisms
RNA therapeutics should not be treated as one interchangeable technology. Small interfering RNA, or siRNA, can reduce production of a selected protein by promoting the breakdown of its messenger RNA. Antisense oligonucleotides can alter RNA processing or reduce gene expression, depending on their design. Messenger RNA products supply instructions for producing a protein. The direction of the intervention therefore matters: reducing a harmful protein and supplying instructions for a useful one are different scientific proposals. The guide to RNA therapeutics compares these mechanisms and the evidence needed to assess their potential.
Alnylam’s Onpattro, or patisiran, shows how mechanism and delivery fit together. Its siRNA promotes degradation of transthyretin messenger RNA, reducing production of the corresponding protein. The medicine is formulated as a lipid complex for delivery to liver cells. Both features appear in the Onpattro prescribing information. For research purposes, that example supports a disciplined question: does another candidate have evidence for its own delivery destination? Success reaching liver cells does not establish that a different formulation will work in another tissue. Evaluate the delivery system and target together.
AI is a discovery method, not a drug modality
Artificial intelligence can help researchers select compounds for testing, but the resulting medicine still belongs to a treatment category such as small molecules or antibodies. A model’s prediction is an experimental starting point. In the 2023 primary study reporting the discovery of abaucin, researchers trained a model using bacterial growth data, tested selected compounds and investigated the resulting antibacterial activity. The work included activity in a mouse wound model. Its value rested on the combination of computational selection and experimental evidence. Those findings did not establish efficacy or safety in human patients.
Assess claims of faster discovery against a defined task and a suitable comparison. “Faster” might refer to identifying an initial active compound, improving its properties or reaching a development milestone. Those achievements should not be presented as equivalent. Ask what work was completed, how performance was measured and whether the comparison included unsuccessful attempts. For an individual candidate, examine the experimental evidence regardless of how it was selected. The guide to AI in drug discovery focuses on separating measurable progress from claims that stretch beyond the reported results.
Separate biological activity from a usable medicine
Scientific assessment becomes clearer when each result is assigned to the question it actually answers. A binding experiment can support interaction with a target. A change in a blood marker can support biological activity. Neither observation, by itself, establishes that patients feel better, function better or live longer. When reviewing a presentation, identify the steps between the measured effect and the proposed patient benefit. Then mark which steps have evidence. This prevents a common analytical mistake: treating an early result as if it answered every later question. It also makes gaps easier to describe without dismissing useful research.
Production deserves the same discipline. Ask whether the material used in early studies represents what the company intends to manufacture for later trials and commercial supply. If the process changes, look for evidence supporting the relationship between the earlier and later products. Investors do not need to design the factory. They do need to recognise when the investment case assumes reliable production that has yet to be demonstrated. The guide to biotech manufacturing examines scale, quality control and production costs. Keep those questions alongside the biological rationale rather than postponing them until a launch forecast appears.
Build a scientific evidence sheet for each candidate
A short evidence sheet makes comparisons more useful than a catalogue of technology labels. Write one sentence describing the proposed mechanism, then record the strongest supporting experiment and its boundary. Keep established findings separate from management’s expectations. A hypothetical entry might read: “The compound reduces the intended protein in treated animals; effective exposure and patient benefit in humans remain untested.” That wording preserves the result without promoting it into a clinical claim. Update the sheet when new evidence answers an open question, contradicts an assumption or changes what the next study needs to establish.
- Biological rationale: What connects the target to the disease, and what alternative explanation could fit the evidence?
- Candidate activity: What has this treatment demonstrated, in which model or patient population?
- Delivery and safety: What evidence supports reaching the required cells at an acceptable dose?
- Next test: Which forthcoming result would strengthen the proposed mechanism, and which result would require revising it?