Research journey
5-MeO-DMT: 17 trials, 58 papers, one canvas
From a 1936 synthesis to a $2.8bn acquisition. Every clinical programme, landmark paper and capital event on one canvas, including the arms that stopped.
Milestones are curated from the Blossom research database. Registry completion dates are shown as planned; probability-weighted forecasts are part of Blossom Enterprise.
Hover any marker to preview its callout on the canvas, then click to pin it. Pinned is what stays open, fills this panel with the full context and links, and is what an export captures. Landmarks stay labelled on the canvas so the exported image reads on its own.
Download the artifact
Exports rasterise the live SVG, so whatever you have toggled is what you get.
Origins
1936 to 2010Synthesised, then shelved for 74 years.1936
First synthesis
Hoshino and Shimodaira synthesise 5-MeO-DMT, 83 years before the first clinical trial.
Toshio Hoshino and Kenya Shimodaira first synthesise the molecule in 1936. It then sits almost entirely outside medicine for the better part of a century. The compound exists, but the apparatus needed to study it in patients (ethics frameworks, trial registries, commercial sponsors) does not yet reach it.
That gap is the single most striking feature of this canvas. Eighty-three years pass between the synthesis and the first person receiving the compound in a registered trial. Nothing on the left half of this timeline is a research programme; it is a molecule waiting for a regulatory environment that will take it seriously.
1959
Isolated from Anadenanthera peregrina
Identified in the seeds used to prepare yopo snuff, connecting the molecule to Indigenous practice.
Isolation from Anadenanthera peregrina (the tree whose seeds are ground into yopo snuff) links the laboratory compound to a far older tradition of use. The molecule was not new to human experience in 1959. It was new to chemistry.
That lineage matters for how the field discusses 5-MeO-DMT now. The programmes further down this canvas are not discovering a substance so much as industrialising one, and the question of who benefits from that industrialisation runs underneath every capital event on the bottom lane.
Evidence builds
2010 to 2018Pharmacology and mechanism, at last.2010-10
Current Drug Metabolism
Pharmacokinetics review
A metabolism and drug-interaction synthesis, the earliest 5-MeO-DMT paper indexed in Blossom.
Published in Current Drug Metabolism, this review pulls together metabolism, pharmacokinetics (how the body absorbs and clears a drug) and drug-interaction data in one place. It is the practical starting point for anyone designing a dosing protocol.
It is also the oldest record in our 59-paper set for this compound, which is worth sitting with. Serious pharmacology arrives 74 years after the synthesis and only nine years before the first trial. The evidence base underneath the clinical era is thinner and younger than the 1936 start date suggests.
2011-01
ended earlyUS Schedule I
The DEA places 5-MeO-DMT in Schedule I, formalising the research barrier.
Scheduling in January 2011 followed a notice of proposed rulemaking issued in 2009. The practical effect is procedural rather than absolute. Every clinical programme on this canvas begins after this date, so Schedule I did not stop the work.
What it did was raise the floor. Each sponsor now carries licensing, storage and handling obligations that a Schedule IV compound would not attract, which favours well-capitalised companies over academic groups. Read the lanes below with that in mind: the two commercial programmes move faster than the academic ones, and the regulatory burden is part of why.
2016-02
Neuropharmacology
Cortical activity disrupted
Preclinical work shows the compound disrupting cortical activity, giving the clinical case a mechanism.
One of the preclinical results (animal and cell studies, before any human trial) that gave the later programmes a mechanistic story to tell. Findings like this are what let a sponsor argue that a compound whose acute effects last minutes could produce change that lasts months.
That argument is doing a lot of work. 5-MeO-DMT is unusually short-acting, which is commercially attractive because it shortens the clinical session and the staffing cost that comes with it. Whether brief exposure delivers durable benefit is still the open question the Phase 3 readout on the right of this canvas is meant to answer.
2017-10
Scientific Reports
Human cerebral organoids
Proteomic changes in human cerebral organoids after 5-MeO-DMT exposure.
A human-tissue model rather than a rodent one. Cerebral organoids (small clusters of lab-grown human brain cells) let researchers watch protein-level changes in human tissue without dosing a person.
Work like this bridges animal pharmacology and the first-in-human studies that open two years later. It is also a reminder of how compressed this section of the canvas is: the three papers here sit within seven years of each other, and the first clinical trial follows almost immediately.
Into the clinic
2019 to 2025Four programmes, two terminations, one withdrawal.2019-03
GH Research
First clinical trial
GH Research opens a Phase I healthy-volunteer study of inhaled GH001.
This is the moment 5-MeO-DMT enters formal clinical development, 83 years after synthesis. A seven-month Phase I in healthy volunteers (people without the condition, dosed to establish safety rather than benefit).
Every later programme on this canvas descends from this point. GH Research reaches the clinic first and stays ahead on paper for years, which makes what happens next instructive: being first did not translate into being acquired.
Worth noting what a Phase I does and does not show. It establishes tolerability and dosing. It says nothing about whether the compound treats depression.
2019-11
GH Research
Phase I/II in depression
The first patient study of GH001, running nearly two years through the pandemic.
The first time the compound is given to patients rather than healthy volunteers, which is the step where a programme starts generating evidence about benefit rather than safety alone.
The near-two-year run reflects pandemic recruitment conditions as much as protocol design. Read trial durations on this canvas with that caveat: the bars measure elapsed time, not effort or difficulty, and 2020 stretched almost everything.
2022-02
Beckley Psytech
BPL-003 first-in-human
Beckley Psytech opens a single-ascending-dose study of intranasal BPL-003.
A different route of administration (intranasal rather than inhaled) and the start of the thread that ends in a $2.8bn acquisition four years later.
Route matters more than it sounds. It shapes how fast the compound reaches the brain, how tightly a clinician can control the dose, and how much of the intellectual property is defensible. Two sponsors pursuing the same molecule by different routes is a patent strategy as much as a clinical one.
2022-12
Usona Institute
Usona intramuscular PK
A non-profit sponsor characterises intramuscular 5-MeO-DMT pharmacokinetics.
Usona is a non-profit medical research organisation, and its presence on this lane is a reminder that not all the clinical evidence here is commercially owned.
Intramuscular administration is a third route, alongside GH001 inhaled and BPL-003 intranasal. Pharmacokinetic work from a non-profit tends to be published in full rather than held as competitive information, which makes it disproportionately useful to everyone else building on this compound.
2023-02
AtaiBeckley
Phase II TRD (still recruiting)
The open-label study underpinning the Phase 3 case, running to late 2026.
Still recruiting today. This is the evidence base Lilly bought into: an open-label study (everyone knows who receives the drug, so there is no placebo comparison) of safety and pharmacodynamics in treatment-resistant depression.
Its registry completion date sits four months out, which puts a hard checkpoint immediately after the acquisition closes. Open-label results are the weakest form of clinical evidence, because expectation alone can move depression scores. That is not a criticism of the study, which was designed to establish safety and dosing.
It is a caution about what a $2.8bn valuation rests on, and the contrast is available on this same canvas: a randomised, placebo-controlled trial of the rival compound published in JAMA Psychiatry four months before the deal.
2023-03
GH Research
ended earlyPostpartum depression (terminated)
A Phase II in postpartum depression is terminated. The results were still published in 2026.
Terminated trials usually vanish. Nobody publishes them, so the field never learns what went wrong, and the same idea gets tried again by someone else.
This one did not vanish. The Phase 2a results were published in the Journal of Clinical Psychiatry in June 2026, three years after the trial stopped. Both events sit on this canvas because both are true, and a timeline that showed only the successes would be marketing rather than evidence.
Termination is also not the same as failure. Trials stop for recruitment problems, funding, or strategic reprioritisation as often as for safety or futility, and the registry record rarely says which.
2023-04
GH Research
ended earlyBipolar II (terminated)
A second Phase II indication stops in the same year it starts.
Two terminations within five weeks of each other narrowed GH Research back toward its treatment-resistant depression core.
That kind of contraction is common and rarely announced as such. What the canvas shows is a sponsor testing three indications in early 2023 and holding one by the end of it. The surviving programme is the one that reads out below.
2023-05
GH Research
Phase II TRD completes
The anchor study in treatment-resistant depression reads out.
With the postpartum and bipolar arms stopped, this becomes the study GH Research’s case rests on. It is also the closest comparator for the BPL-003 programme running one lane below, which matters for reading the acquisition that follows.
Two sponsors, the same compound, different routes of administration and different corporate outcomes. GH Research completed its Phase II first, and went on to publish the only randomised placebo-controlled trial on this canvas. AtaiBeckley was the one Lilly bought.
2023-05
AtaiBeckley
Alcohol use disorder
A single-dose open-label study widens the indication surface beyond depression.
Evidence that the sponsor treated BPL-003 as a platform rather than a single-indication asset, which is relevant to how the acquisition was valued.
Alcohol use disorder is a large market with few effective pharmacological options, so an early signal there raises the ceiling on a compound whose lead indication is crowded. The study is small and open-label, so it establishes feasibility rather than efficacy.
2024-01
atai Life Sciences
atai partners with Beckley
atai Life Sciences takes a stake and joins forces on BPL-003 development.
The first consolidation step on the BPL-003 thread. Capital and development capability arrive well before the clinical case is complete, which is the normal sequence in biotech and worth stating plainly: money commits to this compound in 2024, on the strength of a first-in-human study and one open-label trial still recruiting.
atai brought a portfolio approach, spreading risk across several compounds. Beckley brought the asset.
2024-05
Nature
Structural pharmacology in Nature
Receptor-level structures for 5-methoxytryptamines, published in Nature.
The highest-profile venue any 5-MeO-DMT work has reached. Structural pharmacology maps how the molecule actually binds its receptor, which is what lets chemists design analogues that keep the antidepressant effect and drop the hallucinogenic one.
That matters commercially as much as scientifically. Every sponsor on this canvas is eventually asked whether the psychedelic experience is necessary for the benefit, and structural work is where an answer would start.
2024-10
Biomind Labs
Sublingual microdosing
Two Biomind Labs Phase I/II studies test a sublingual, low-dose route.
A fourth route of administration and a different dosing philosophy from the single high-dose sessions that define the GH001 and BPL-003 programmes.
Microdosing (repeated sub-perceptual doses rather than one intense session) targets a different commercial model: no clinical supervision, no session staffing, far lower delivery cost. The evidence base for microdosing across psychedelics is weak, and placebo-controlled studies have repeatedly failed to separate low doses from placebo. Two small Phase I/II studies do not change that.
2024-12
JAMA Psychiatry
Adverse events meta-analysis
A systematic review of adverse events across classic psychedelics, in JAMA Psychiatry.
The safety counterweight to everything else on this canvas. A systematic review pools adverse events across classic psychedelic trials, which is the only way to see signals too rare for any single study to catch.
Timelines built from company announcements tend to omit this category entirely. It sits on the science lane because safety evidence is evidence, and because the acquisition three lanes down is partly a bet that the safety profile holds at scale.
2025-03
Johns Hopkins University
ended earlyJohns Hopkins study withdrawn
The Trifecta research study is withdrawn before enrolling anyone.
Withdrawn is a specific registry status: the study was registered and then closed without enrolling a single participant.
Records like this are easy to quietly drop from a pipeline chart, and most commercial trackers do exactly that. Leaving it visible keeps the academic lane honest and shows something real about the period, which is that registered intent and delivered research are not the same quantity.
2025-11
AtaiBeckley
AtaiBeckley formed
The combination completes; the merged company trades on Nasdaq as ATAI.
Two years after the partnership, the companies combine fully and the merged entity trades on Nasdaq as ATAI. This is the second custody change on the BPL-003 lane.
Eight months later that entity is acquired. Anyone reading the lane from left to right can watch a single asset pass through three owners in under five years, while the underlying clinical evidence advances by one open-label study.
2025-12
Addiction
Alcohol use disorder results published
The open-label Phase 2 proof-of-concept in alcohol use disorder reports in Addiction.
The 2023 alcohol study reads out in print two years after it closed. Proof-of-concept and open-label, so it establishes feasibility rather than efficacy, but it moves the second indication from claim to published record.
Publication lag is worth noticing across this canvas. Trials finish long before anyone outside the sponsor can read the numbers, which means a valuation can move on data the field has not yet seen.
Capital arrives
2026 onwardsBig pharma buys the option.2026-02
J Psychopharmacology
BPL-003 Phase 2 results published
The uncontrolled, open-label Phase 2 in treatment-resistant depression reports in print.
Five months before the acquisition, the BPL-003 depression data reach the literature. The title states the limitation plainly: uncontrolled and open-label, so there is no placebo arm and no blinding.
This is the published evidence base Lilly bought. It is a real result and a weak design at the same time, and holding both of those in view is the whole point of putting publications on the same canvas as deals.
2026-03
JAMA Psychiatry
GH001 randomised trial in JAMA Psychiatry
GH001 beats placebo in a randomised controlled trial, the strongest evidence on this canvas.
The first randomised, placebo-controlled evidence for any 5-MeO-DMT compound, published in JAMA Psychiatry in March 2026. Randomisation and a placebo arm are what separate a suggestive result from a credible one, and until this point every study on this canvas lacked both.
It belongs to GH Research, not to the programme Lilly acquired. Read the two clinical lanes together and an awkward pattern appears: the sponsor with the stronger published evidence is not the sponsor that got bought.
One caution before treating this as settled. A single randomised trial is a beginning, not a conclusion, and independent replication is what turns it into an approvable claim.
2026-03
University of Basel
Basel IV study opens
University of Basel begins a Phase I of intravenous 5-MeO-DMT, running to 2028.
Academic pharmacology continues in parallel with the commercial programmes. Intravenous dosing gives the tightest possible control over how much compound reaches the bloodstream and when, which makes it the cleanest setting for measuring acute effects.
This is also one of the few studies on this canvas that runs past the pivotal readout. Whatever the Phase 3 shows in 2029, independent characterisation of the compound will still be underway in Basel, and that work is not owned by anyone with a commercial position in the outcome.
2026-07
Eli Lilly
Eli Lilly acquires AtaiBeckley
$6.75 a share in cash, about $2.8bn upfront and up to roughly $3.8bn including CVRs.
Big pharma buys into 5-MeO-DMT. Lilly pays about $2.8bn upfront, with the rest in contingent value rights (payments that only trigger if named development and regulatory milestones are hit) tied to BPL-003 and VLS-01.
That structure is the most informative part of the deal. Roughly a quarter of the headline price is explicitly staked on events that have not happened, including the pivotal readout sitting three years to the right on this canvas. Lilly is not paying $3.8bn for evidence. It is paying $2.8bn for an option, and the market agreed enough to move the shares 30% on the day.
Which raises the question the whole right-hand side of this timeline is really about: what has to be true in 2029 for this to have been worth it?
2026-11
AtaiBeckley
plannedPhase II TRD completes
Registry-listed completion date for the open-label study.
A public registry date rather than a forecast, and the next hard checkpoint on the acquired programme. It falls within months of the acquisition closing, which makes it the first real test of what Lilly bought.
2028-03
University of Basel
planned2029-03
Eli Lilly
plannedBPL-003 pivotal data expected
The Phase 3 pivotal readout, guided for early 2029, is what the acquisition is priced against.
The furthest point on the canvas, and the one everything to its left is arguing about. This is where 5-MeO-DMT either becomes a medicine or does not.
The date is company guidance rather than a registry-confirmed completion, so it is drawn as planned and should be treated as the optimistic case. Pivotal trials slip routinely.
It is also the first properly controlled test of BPL-003. Every earlier study on this thread is open-label or early-phase, so the strongest evidence for the acquired asset arrives three years after the cheque cleared. GH001 already has its randomised result, two lanes up and three years earlier.
2029-09
plannedenterpriseProbability-weighted approval scenarios
Filing timing, approval probability, and comparator readouts across the 5-MeO-DMT field.
Everything else on this canvas is public record: registry entries, published papers, filed announcements. This layer is not.
Blossom Enterprise models filing timing and approval probability against comparator programmes, which is the analysis that turns a timeline into a decision. It is the one marker here you cannot verify from a registry, and it is deliberately the only thing behind a paywall.
5-MeO-DMT
Mebufotenin: the research journey