Oncology · Open data · Drug discovery

OncoSignal

An open oncology intelligence tracker. For one cancer at a time it shows what moved in the trial registry, who is running the trials, which targets have human genetics behind them, and what was presented at the 2026 meetings. Every row links to its source.

Open OncoSignal See the lung cancer dashboard See a target page: EGFR

Where it started

OncoSignal was inspired by Khushi Shelat's clinical trial research demo for Parallel. Her demo showed how much of a trial landscape you can assemble from public sources. I wanted the same thing kept current for oncology, one cancer per page, with the genetic evidence for each target beside the pipeline.

The reason to build it is in my essay Curing All Disease, Ahead of Schedule. It argues that drug development is held back less by discovery than by herding: companies crowd onto the same validated targets. OncoSignal puts that herd on one screen, then asks where the human genetic evidence points that the pipeline has not gone yet.

I'm Ashish Makani (tp53; GitHub and X: inventcures), an academic in India working on computational protein and antibody design and drug discovery.

The aim

Surface what has been tacit

What can we uniquely surface to our target users, which so far has been tacit, and living in expert drug hunters' heads and mental models?

Show some insights, learnings, which no other comparable platform currently does, to our target users. That is the bar for what OncoSignal builds next.

On the live site

What is there today

Counts as shown on the live site on 11 November 2026. The site is still being built, so they will have moved by the time you read this.

18
cancers, each with its own dashboard: solid tumours, blood cancers, brain tumours and nine childhood cancers.
285
2026 meeting abstracts found by web research and verified against a source page: ESMO 146, ASH 98, SITC 18, ASCO 13, AACR 10.
5
views per cancer: trial news, landscape, lead investigators, targets and abstracts.
4
public data sources behind every view: ClinicalTrials.gov, PubMed, Open Targets and Crossref.

What it shows

Five views for each cancer

The screenshots below are of the live site, taken on 11 November 2026. Click one to open that view.

1. Trial news

Events derived from changes in the registry: an asset's first Phase 2 or Phase 3 trial, new pivotal trials, results posted, readouts coming due, and stopped trials with the reason the registry gives. New trial papers from PubMed sit in the same feed.

OncoSignal trial news for non-small cell lung cancer: counts of phase advances, new Phase 2+ trials, results posted, readouts due and trial papers, above a dated feed of stopped trials, posted results and papers.
Trial news for non-small cell lung cancer.

2. Landscape

Every asset's trials on one timeline, sized by phase, so you can see who moved from Phase 1 to Phase 3 and when.

OncoSignal landscape for non-small cell lung cancer: one row per drug with its lead sponsor and a dot for each trial on a 2009 to 2026 timeline; Phase 3 trials are orange.
The lung cancer landscape. Each row is an asset; each dot is one of its trials; orange is Phase 3.

3. Lead investigators

Principal investigators and study chairs named on the registry, ranked by Phase 2 and later trials in that cancer, with their trials and PubMed output.

4. Genetically supported targets

Targets from the Open Targets Platform, with inherited (germline) and tumour-mutation (somatic) evidence split out. A chart sets genetic evidence against drug evidence, so the targets with strong genetics and few drugs stand apart.

OncoSignal targets chart for non-small cell lung cancer: a scatter plot of human genetic evidence against drug evidence for the top 50 Open Targets targets, with TP53, TERT, DICER1 and PIK3CA in the strong genetics, few drugs corner and KRAS, ERBB2 and KIF5B in the corner with drugs in the clinic.
Lung cancer targets: genetic evidence (up) against drug evidence (right). Scores are Open Targets' summary of how much evidence exists, not the chance a drug will work.

Each target has its own page, with the drugs recorded against it by type and stage, and the evidence for it across cancers.

OncoSignal target page for EGFR: a flow chart from drug type (small molecule, antibody, antibody drug conjugate) to furthest stage reached (approved, Phase 3, Phase 2, Phase 1, withdrawn), and a dot plot of inherited and tumour-mutation evidence for EGFR in fifteen cancers.
The EGFR target page: drugs by type and furthest stage, then germline and somatic evidence by cancer.

5. 2026 meeting abstracts

Abstracts from ESMO, ASCO, AACR, ASH and SITC 2026, filed by cancer, late-breakers first, each with its trial, NCT ID and source link. ASCO, AACR and SITC supplements are searched live through Crossref. ESMO and ASH abstracts are pulled with the Parallel API, and a row is kept only if its number and title were verified against a fetched source. Abstract text is never stored.

OncoSignal abstracts view for non-small cell lung cancer with tabs for SITC, ESMO, ASCO, AACR and ASH 2026, listing ESMO late-breaking abstracts with presenter, trial name, NCT ID, drugs and a verification note.
ESMO 2026 lung cancer abstracts, with how each row was verified.

Ground rules

Real data only

  1. No mock, sample or synthetic data.Not in the app and not in its code.
  2. Every row links back to its source record.A trial to ClinicalTrials.gov, a paper to PubMed, a target to Open Targets, an abstract to its DOI or source page.
  3. If a source fails, the view shows the error.It never fills the gap with substitute data.
  4. Unverified abstracts are dropped.Web research finds candidates; only rows confirmed on a source page stay.

Sources

Where the data comes from

ViewSource
Trial news, landscape, investigatorsClinicalTrials.gov API v2
Papers, investigator outputPubMed E-utilities
TargetsOpen Targets Platform (EMBL-EBI)
ASCO, AACR and SITC abstractsCrossref supplement DOIs
ESMO and ASH abstractsParallel web research, each row verified against a source page
Paediatric target flagFDA Pediatric Relevant Molecular Target List

Full details, including how each event is derived, are on the app's Sources & roadmap page. JSON for every view is at /api/i/<indication>/<view>.

Where it stands

Built versus planned

  1. Live

    Now

    Live registry, PubMed, Open Targets and Crossref data on every view; verified ESMO and ASH abstracts; eighteen cancers and a paediatric view.

  2. Next

    Complete abstract books

    Full harvests for ASCO, AACR and SITC, session-by-session pulls for ESMO and ASH, and topline results as structured fields.

  3. Next

    Readouts and deals

    Which trial read out and what was reported, from company disclosures, with citations, joined to NCT IDs.

  4. Later

    Briefs, alerts and a herd map

    A weekly brief per cancer, alerts on phase advances, and a map of pipeline crowding against genetic support.

Not medical advice. OncoSignal is a research tool built on public records. It is not a guide to treatment.

More of my work on trials and oncology

Curing All Disease, Ahead of Schedule →The essay behind OncoSignal. FDA CRL analysis →An analysis of FDA Complete Response Letters. AACR 2025 clinical trials dashboard →Trial phases, therapeutic targets and outcomes from AACR 2025. ASCO 2025 data visualization →Oncology research presented at the ASCO 2025 annual meeting.