Finding biotech early, from the public record.
Diaspore is an AI business development and venture capital team. It maps who matters in a field, ranks the labs and young companies to watch, and reads one company's science, filings and trials the way a deal team would, with a dated record of every call.
Use it here
Use it from your own assistant
Connect your AI assistant with one address and ask in your own words.
Could it have spotted them early?
The test
8 of the top 25
young companies the model picked on 2022-12-31 went on to raise a seed round or larger within 24 months. Picking at random would find 1.7; the closest comparison, an AI given the same facts, found 7.
The model was fixed first and scored once, on a date it had never seen. 75 of the 1,097 companies in the test raised.
On the same companies
| Picking at random | 1.7 of 25 |
| The best simple rule | 5 of 25 |
| A hand-set scorecard | 6 of 25 |
| An AI given the same facts | 7 of 25 |
Across the whole list the model and the AI are level: the test cannot tell them apart.
One it found
What was public on 2022-12-31
Basking Biosciences, Inc.
| Financing filed with the SEC so far | $1.9 million |
| Most recent financing filing | 2020-09-10 |
| Federal small-business research awards | 2 |
| Year of incorporation | 2019 |
What the model said, and what happened
10 of 1,097 young companies, ranked on that day's public record
On 2023-11-15 the company filed a financing of $55.0 million.
Chosen by a rule, not by hand: of the model's top 25 picks that went on to raise, the one with the largest next round. Most of the top picks did not raise.
After a find: the rest of the team's work
Diaspore is an AI business development and venture capital team. Finding a company is the first job. The same system then reads it, places it among earlier deals, sets out the decision and keeps the record. Each job below shows a line of its real output: the rank comes from the seed-stage test above, the rest from one worked run.
Source
Find who has public evidence on a topic or a company as of a date, and how busy the field already is.
In the registry, 35 other sponsors had 82 trials in the company's conditions started on or before 2015-03-02.
Screen
Rank labs and young companies on the public record of that day.
Basking Biosciences, Inc. was ranked 10 of 1,097 young companies on the public record of 2022-12-31, before it filed its next financing.
Diligence
Read the company's own filings and the trial registry, and quote what they say about the science, the competition and the cash.
The company states that Fovista binds and inhibits PDGF, which it says strips pericytes from newly formed blood vessels.
The company reports cash, cash equivalents and available for sale securities of $463.6 million as of December 31, 2014.
Value
Place the company among earlier deals at the same stage, as a range with its assumptions.
Among 4 earlier acquisitions of companies at the same clinical stage, 2 have a disclosed value; they span $122.2 million to $255.7 million.
Decide
Set out the case for, the case against, and what would change the read. No verdict and no new number.
The lead program has a stated mechanism and has reached the latest trial phase in the registry, with a top-line readout expected on a defined timeline.
Many rival anti-PDGF and related candidates are named by the company itself, including a large pharma co-formulation with the current standard of care, which crowds the lead program's space.
Track
Fix each call before its outcome exists, then check it when it comes due.
A full run on another company was fixed in advance and will be checked on 2028-03-31.
One company, start to finish
Ophthotech Corp., as the public record stood on 2015-03-02
The team read the company's annual report and the trial registry as they stood that day and wrote a report of sourced claims. A second AI that had not seen the run checked each claim against its source: 30 of 31 matched, and it caught all 5 of the wrong claims slipped in to test it.
Everything in one place
Results
What it has shown so far, each result beside chance, a simple rule and an AI given the same facts.
Run
One company as of a date, start to finish, every claim linked to its source.
How it works
The six jobs of a deal team and what Diaspore produces for each.
Find
Search the science as of any date, map a field by lab and institution, and test your own picks.
Targets
Who is publishing on each drug target, and how fast each field is filling up.
Rank
Labs and companies to watch, with the reasons behind each rank.
Connect
Use all of it from your own AI assistant.
Track record
The calls fixed before their outcomes exist, and when each comes due.
What it is built from
Public data only: papers, grants, patents, trial records and SEC filings. How it was tested.