Summer 2026 Product Roundup
DrugBank's knowledge graph got a lot more connected this quarter. New entry points, deeper target biology, and richer trial data now connect end to end through MCP, giving biopharma teams a complete picture with fewer blind spots.
This quarter, we made our knowledge graph much denser. We added new entry points, deeper connections between targets and the biology around them, clearer visibility into the competitive landscape, and structured detail inside trials. Each addition opened another path through the graph, so questions that used to take multiple separate lookups now resolve in a single query, with fewer blind spots along the way.
All of this has also been made available through the DrugBank MCP, bringing DrugBank directly into your environment, so it runs alongside your data and the tools you already use.
To show what these new connections enable, we'll follow one target, JAK1, across the knowledge graph. From a single target you can understand the biology it sits in, the drugs that target it, who's developing those drugs, and how their trials compare across diseases, drilling into eligibility, outcomes, and adverse events.
More ways into the knowledge graph
A research question can begin anywhere: a target, a compound you're profiling, a protein sequence you're chasing. We added new ways in, so you can start from what you already have and follow the same connected path into the full landscape around it.
Search from a chemical structure
Start from a SMILES string, InChI, or DrugBank Drug ID to receive ranked drug matches, then refine by weight, similarity threshold, drug type, and more. Now as an entry point in our MCP, use a compound to pull what’s known about structurally similar drugs, their targets, indications, trials, and sponsors.
Search from a protein sequence
An amino acid or nucleotide sequence returns ranked target matches by similarity, connected to the greater drug context. Start from a FASTA sequence, UniProt accession, DrugBank Protein ID, or gene symbol to see if close relatives are already drugged, showing if you’re working in novel biology or already-explored territory.
Which drug targets are most similar to JAK1 by protein sequence, and what approved drugs hit each one?
JAK1's closest sequence relatives are the other Janus kinases, and they share most of the same inhibitors. FRK is a distant match with only a multikinase drug.
Deeper biological foundation around targets
A target is now a hub to explore the biology around it. From one target you can reach the pathways it sits in, the drugs that act on it, the trials those drugs run in, and each drug's pharmacology, so the mechanism and the landscape stay connected.
Connect a target to the trials testing it
Targets now link to clinical trials through the investigational drugs that act on them, so a target opens into the clinical pipeline. With greater clinical target coverage, you can surface hotspots where development is crowded, identify gaps where no one is developing yet for potential whitespace, and assess your competitive position around a target of interest.
Map a target to the pathways it sits in
We added Reactome as a second pathway source alongside SMPDB, so most proteins now sit in at least one pathway. Find target whitespace at the pathway level, even in crowded diseases, or trace how a mechanism propagates and compare related targets along the same pathway.
Compare candidates on structured pharmacology
ADME, PK, and PD details are now queryable, machine-readable fields in the MCP. Compare exposure, half-life, clearance, and protein binding across approved drugs to inform dosing and differentiation.
Beyond JAK1, which other targets in the same pathways are already drugged, and by what?
Every Janus kinase has approved inhibitors, largely the same multi-target molecules. No STAT has an approved drug yet, though STAT3 is heavily pursued in development.
Read the competitive landscape and the science surrounding it
Corporate M&A activity can hide who actually holds an asset, and two drugs with the same headline indication can compete on very different terms. From a target you can now trace the drugs developed against it, the sponsors behind them mapped to the parent company, how far each program has progressed, and the mechanisms, indications, and interactions that set them apart.
Trace every trial sponsor to its parent company
Company normalization now covers all industry sponsors in DrugBank, bringing trials filed under subsidiaries, historical names, and rebranded entities under a single company. You can now see companies’ true trial footprint with a complete read on company and asset ownership across the industry.
Tell similar drugs apart by how they're used
Two drugs that look identical at the indication level can now be compared on the specific terms they compete. Structured indications can now be queried through the MCP to reveal the detail behind a headline condition, including on-label and off-label use, dose, region, patient population, combination use, and investigational versus indicated status.
Evaluate a drug’s interaction burden
Drug-drug interactions are now searchable from a single drug, so you can screen a candidate's interaction profile and weigh its burden against the approved comparator. In diligence, that interaction load becomes a differentiation and risk signal alongside efficacy and mechanism.
Which companies are developing JAK1 inhibitors, and what indications are each pursuing?
Eleven approved JAK1 inhibitors and their lead developers. Development crowds into rheumatoid arthritis and atopic dermatitis, alongside a smaller, largely owned myeloproliferative and oncology camp.
Understand what each trial found
We added what a trial set out to measure and what it reported, including adverse events, so you can compare endpoints across a class and scan safety signals across an indication.
Compare what trials set out to measure
Outcome measures capture the primary, secondary, and other endpoints a trial was designed to prove, now consolidated and searchable. A clinical trial can reveal valuable signals beyond its primary outcome; you can now investigate secondary outcome measures and patient subgroups to identify indications and patient populations that deserve a closer look.
Surface trial results and adverse events
Trial results include structured adverse events, which are linked to DrugBank condition ID, connecting them to other disease ontologies like MedDRA, MeSH, and SNOMED. Because adverse event data is structured, you can trace their patterns across disease subtypes. Paired with clinical trial ‘why stopped’ reasons, you can scan terminated trials for safety patterns to distinguish design flaws from a broader safety signal.
Filter trials by who they enrolled
Eligibility criteria let you narrow a trial landscape by the population a study enrolled, down to details like age, disease subtype, and prior treatment. Combined with outcomes and adverse events, you can compare trials that enrolled comparable patients, so differences in results reflect the drug rather than the population.
Which adverse events were most common across JAK1 inhibitor trials in atopic dermatitis?
| # | Adverse event | Participants | Rate | Seen on |
|---|---|---|---|---|
| 1 | Nasopharyngitis | 209 | 10.3%of 2022 | Oral and topical |
| 2 | Upper respiratory tract infection | 163 | 7.1%of 2292 | Oral and topical |
| 3 | Acne | 135 | 10.3%of 1313 | Oral and topical |
| 4 | Dermatitis atopic | 120 | 7.1%of 1693 | Oral and topical |
| 5 | Headache | 101 | 6.0%of 1693 | Oral and topical |
| 6 | Nausea | 79 | 12.7%of 623 | Oral only |
| 7 | Blood creatine phosphokinase increased | 29 | 7.0%of 412 | Oral only |
| 8 | Folliculitis | 24 | 5.8%of 412 | Oral only |
| 9 | Urinary tract infection | 21 | 5.1%of 412 | Oral only |
| 10 | Oral herpes | 17 | 5.0%of 342 | Oral only |
The ten most frequently reported events across seven trials and 2,292 participants on a JAK1 inhibitor. Nasopharyngitis and upper respiratory tract infection appear on every form; acne, raised creatine phosphokinase, folliculitis, urinary tract infection, and oral herpes are almost entirely oral.
In summary
Each addition this quarter was a piece of a much larger puzzle. In isolation, these changes are helpful, but working together as a whole they create a meaningful shift in how quickly and accurately you can move through your work.
A new entry point only matters because of what it connects to, and a compound search is worth more now because it reaches the targets, sponsors, and trials behind it, not because compound search exists on its own. A denser graph doesn't just add features one at a time. It multiplies the paths between the ones you already had, so each new connection makes every other connection more valuable. That's the real shift this quarter, not a series of individual additions, but a single more connected, complete graph.
The JAK1 walkthrough followed one target, and that same path now runs from any target, compound, or protein sequence you start with.
What’s next
Looking ahead, expect expanded pre-clinical coverage, more of DrugBank inside the MCP, and new tools for navigating the biomedical and competitive landscape.