Target Atlas for Smarter Target Prioritisation for Parkinson’s

Target prioritisation decides which drug candidates ever reach patients, and Phylo’s Target Atlas now automates much of it. Built on Biomni Lab, the platform ranks thousands of targets with source-linked evidence and offers a free Parkinson’s edition developed with the Michael J. Fox Foundation.

Image

Why Does Target Choice Decide Drug Discovery Odds?

Every drug programme starts with a bet on a biological target. If that bet is wrong, no trial design or budget can save the project.

Target selection is therefore one of the most consequential decisions in Drug Discovery, yet it often rests on the thinnest evidence.

Human genetics offers one of the clearest sources of data for drug discovery.

Few signals in early discovery carry that much weight. A recent Nature analysis of 13,022 target and indication pairs found that mechanisms with human genetic support had an estimated 2.6-fold higher probability of success from Phase 1 to approval.

Genetics is only one input, though. Teams must also analyse gene expression, safety signals, tractability, biomarkers, and competitor activity.

Gathering all of that across thousands of targets and diseases takes far more time than manual review can sustain. New research and clinical evidence can change the picture before the lab work is finished.

And the cost of error is steep. Programmes that pick a weak target can run for years before the failure shows, and by then R&D teams have committed laboratory capacity, clinical budgets, and partner goodwill.

Stronger evidence upstream gives leaders a firmer base for portfolio decisions.

Phylo built Target Atlas to close that gap in target prioritisation.

The Target Atlas platform assembles evidence continuously, so teams review current, auditable data instead of a snapshot that ages by the month.

Better evidence at the point of selection protects every later investment.

View interview How Should Pharma Companies Audit AI Vendors?

How Should Pharma Companies Audit AI Vendors?

“This is one of the riskiest industries there is because you're dealing with human lives, you're dealing with experimental protocols, and you're dealing with regulatory bodies where you might not get another shot at that clinical trial."

Image

How Does Target Atlas Drive Target Prioritisation?

Target Atlas runs on Biomni Lab, Phylo’s Integrated Biology Environment.

It launches with thousands of target and disease assessments, compiled by hundreds of AI agents.

Automated workflows gather the evidence, organise it, and rank targets across the categories that matter for each disease.

Researchers stay in control. They can chat with existing profiles, add new evidence categories, and upload their own data.

Scores then update to show how each addition shifts the order. Every change is visible, so colleagues can reproduce the reasoning.

Target profile pages

Each assessment opens into a profile page. Sections cover genetics, patient evidence, expression, clinical activity, safety, tractability, competition, and therapeutic hypotheses.

Every section carries its own citations and time stamps, so reviewers get a clear audit trail.

Traceability really matters in drug development. Leaders can see why a target ranks where it does, challenge the inputs, and record the rationale for portfolio reviews.

The AI agents behind the evidence

Three example agents show how profiles are built:

• Single-cell analysis: For diseases with a pinned CELLxGENE dataset, the AI agent runs donor-aware differential expression from raw counts. It labels non-significant results openly.
• Full-text literature review: The AI agent sorts evidence into human genetics, patient samples, in vivo studies, and in vitro studies.
• Competitor checks: The agent reviews each trial drug by drug. That separates genuine competition from gene symbol clashes and biomarker arms.

Separate review agents then audit every profile. They check cited studies and trials against source records, correct misstated findings, and flag anything that is not direct evidence for that target in that disease.

Phylo admits that errors can still slip through and invites community feedback. That candour suits a tool that informs high-stakes decisions.

Interactive 2x2 plots

A ranked table can hide trade-offs between criteria.

Target Atlas lets users plot any two evidence fields against each other, with cohort-median guides on continuous axes. Targets missing either value are omitted rather than scored as zero, which avoids false precision.

The dashboard score is a weighted average of normalised columns. Change any weight, and the whole cohort re-ranks instantly.

Image

Inside the Parkinson’s Edition With MJFF

Parkinson’s disease makes a truly demanding test case.

Current treatments ease symptoms but do not halt progression, and the field hosts many competing biological hypotheses. Choosing which to back is difficult, and a wrong choice costs years of work.

The Michael J. Fox Foundation’s Targets to Therapies initiative tackled target prioritisation with community nominations, expert review, evidence scorecards, and a published method.

Over nine months, it narrowed 229 candidates to 59 for deeper review, as reported in npj Parkinson’s Disease.

The Parkinson’s edition of Target Atlas keeps all 229 candidates explorable. Filters isolate the top 59 and top 21 tiers, and source-linked evidence sits alongside the foundation's framework.

Researchers can compare targets, upload their own data, and test whether it strengthens, challenges, or widens a profile.

Where expert judgement meets molecular data

One 2x2 plot compares expert rank with dopaminergic-neuron expression. Of 229 targets, 215 have values on both axes. Selecting MAPT opens the scoring and source details behind its position.

Where the two measures agree, confidence grows. Where they diverge, a candidate deserves a closer look.

That is how a ranked list becomes a prompt for better research questions.

This pairing is important because expert opinion and molecular data fail in different ways. Experts may favour familiar targets, while expression data may miss disease biology that sits outside one cell type.

Viewing both together exposes the blind spots of each.

Expert judgement stays central. Scientists still frame the questions, weigh the evidence, interpret the results, and validate findings.

The platform simply makes the underlying evidence easier to reach.

View interview AI in Pharma: Why the Future of Healthcare Starts With Patients, Not Tech

AI in Pharma: Why the Future of Healthcare Starts With Patients, Not Tech

Kate O’Reilly, President & Chair of the Healthcare Businesswomen’s Association (HBA) Dublin-Ireland Chapter & Healthcare Transformation Partner at Roche, discusses AI in pharma, patient engagement, and the future of healthcare innovation.

Image

Public and Enterprise Access for Pharma R&D Teams

Phylo offers two routes into the platform. The public edition is free with a Biomni account and includes:

• The full Parkinson’s dashboard, open to everyone, with all data, evidence, and features
• Thousands of target and disease assessments, with a roadmap to cover all major diseases over time
• Evidence-supported target cohorts: A target enters once it clears at least one stated channel, such as human genetics, disease-relevant expression, clinical activity, a curated biomarker, or literature nomination
• Standard dashboards, profiles, and 2x2 plots that highlight evidence gaps and analytical opportunities
• Private extensions that let users add criteria without changing the shared public atlas

The enterprise beta suits R&D teams. It uses an annual licence for a single-tenant Biomni deployment, with usage on a credit-based model.

Teams can assess all 20,000-plus targets per selected disease, including evidence-sparse and previously un-nominated ones.

They can merge proprietary screens and internal omics data with the public evidence layer.

They can also train AI and machine learning models that estimate disease-specific evidence weights alongside expert-set weights.

Target-pair dashboards then support bispecific and dual-target CAR-T strategies, using measured co-expression, AND/OR logic, per-antigen safety, and clinical precedent.

What should R&D teams consider?

First, the Target Atlas launch has been a vendor announcement. It offers no outcome data showing that Atlas-guided choices raise clinical success.

Second, the output is only as good as its sources and agents, and Phylo itself says errors can occur. Lastly, scientists must still own the final judgement.

For pharma, the question is whether to build versus partner.

Large companies often run internal scoring pipelines. Atlas offers a shared public layer that smaller biotech teams can use at no cost, which narrows the evidence gap between large and small players.

Larger groups may then differentiate through proprietary data rather than basic evidence assembly.

Sponsorship may prove the most interesting part. Phylo wants foundations, consortia, and patient organisations to fund further disease editions that stay free to their research communities, with each organisation’s prioritisation framework built into the scoring layer.

That model could spread expert-led frameworks well beyond Parkinson’s.

Image

Turning Target Prioritisation Into Better Decisions

Target Atlas shows how AI can be used for better target prioritisation to create better drug discovery.

Evidence assembly becomes continuous, auditable, and open to scrutiny, while scientists keep the final call. That shift could shorten the path from a promising hypothesis to a well-justified programme.

The platform pulls genetics, expression, safety, tractability, and competitor data into one source-linked view.

Its Parkinson’s edition carries a respected expert framework, with 229 candidates, a 59-target shortlist, and open access for every researcher. Its enterprise tier adds proprietary data, AI scoring, and target-pair dashboards for industry teams.

The open question is outcomes. Until teams publish evidence that platform-guided choices improve trial success, Target Atlas remains a powerful accelerator for evidence review, not a replacement for scientific scrutiny.

The strongest programmes will pair automated evidence with sharp expert questions.

At Pharmatica, we follow the tools, evidence, and strategies that shape early-stage Drug Discovery. Our Insights helps pharma R&D teams judge which innovations deserve investment, so decisions rest on insight rather than hype.

Explore our latest coverage and join the conversation that moves the sector forward.

Pharmatica: Insight. Connection. Impact.