Coverage

Coverage comes from two pipes, and understanding them explains everything else.

• Approved and marketed drugs and diagnostics enter comprehensively, regardless of what disease they treat. This pull is indication-agnostic, so the approved landscape is broadly covered: vaccines, antivirals, antibiotics, metabolic drugs, oncology drugs, and drugs used in aging. It is broad rather than a guaranteed every-single-one, and it excludes the non-drug products noted below.
• Pipeline drugs (still in trials, not yet approved) enter through a set of therapeutic-area searches: oncology, immunology, cardiology, metabolic disease, neurology, rare disease, and psychiatry, including the psychedelic pipeline. The pipeline is deep for these areas and thinner for areas not actively searched.

• Essentially all FDA-approved drugs, across every disease area. Approved drugs are pulled comprehensively from drug databases (ChEMBL and Drugs@FDA), so coverage of approved small molecules and biologics is broad and indication-agnostic. It is not a guaranteed one hundred percent: a handful of older, discontinued, or non-drug approved products (see "what is left out, on purpose") fall outside it.
• Major molecular and genomic diagnostics (liquid biopsy, prenatal and carrier screening, companion diagnostics, and the like). Diagnostic coverage is curated rather than exhaustive: BioCosm tracks clinically meaningful molecular tests, not the full universe of FDA-cleared devices and assays.
• The active clinical pipeline (Phase 1 through 3) for oncology, immunology, cardiology, metabolic disease, neurology, rare disease, and psychiatry.
• For each program: its biological target, mechanism, trials, sponsor, and where available, revenue and a probability-of-success estimate.

A few areas have their approved drugs but a thinner experimental pipeline, because they sit outside the active search set.

• Early-stage infectious disease. Approved antivirals, antibiotics, and vaccines are present; the youngest experimental vaccine and anti-infective programs may not be.
• Longevity and anti-aging. These therapeutics are covered, but filed under the disease their trial targets, not under a single "longevity" heading. Regulators do not recognize aging as an indication, so a reprogramming, NAD-boosting, or senolytic program appears under the specific disease it is tested in (for example an eye disease or a metabolic condition) rather than grouped together. They are on the map; they are simply spread across disease regions.

• Hardware medical devices and lab instruments (analyzers, monitors, pumps, sensors). BioCosm maps drugs and molecular diagnostics, not hardware. A dedicated device view may come later.
• Nutritional supplements, parenteral-nutrition components, and similar non-drug products. These are FDA-regulated but are not drug development programs.
• A small number of programs are redacted where the author has a conflict of interest; those pages show a disclosure in place of analysis.

The number on the map counts distinct drug and diagnostic programs, not raw trial records. A single drug appears in public data under many names and many trial arms: "sotorasib", "AMG 510", and "sotorasib 960mg" are the same molecule across different studies. BioCosm collapses all of those into one node through nightly deduplication.

That is why the count is a curated figure: tens of thousands of trial rows reduce to a few thousand real programs. It is also why the internal database is larger than what the map shows. The difference is duplicates merged and noise removed, so the map stays a clean count of distinct programs rather than an inflated trial tally.