Atom Research — Independent scientific research from Pakistan

§ 01 · Mission

Pakistan is missing from the global scientific record. We are here to change that.

Based in
Pakistan

Built for
global science

An independent scientific research organization based in Pakistan — designing primary human studies and building the datasets and models this population has never had.

~240 million people, almost entirely missing from the datasets that medicines, public-health policy, and human-centered technology are built on. Science cannot represent populations it has never measured.

Underneath the work sits an instrument layer — specialized MRI cohorts, blood biomarker panels, longitudinal behavioral data. On top, foundational models calibrated to Pakistani bodies and brains.

§ 02 · What we study

Four areas of practice. Each at the smallest defensible unit.

Updated
Apr 2026

Active studies
across multiple sites

A · Healthcare

Clinical & translational studies

Clinical and translational studies in partnership with hospitals. Endpoints that travel back to clinicians, not just to journals.

ClinicalPublic healthLongitudinalOutcomes

B · Behavioral science

Behavior in real-world environments

Behavioral science in real-world environments — questions rarely asked of these populations, instrumented to hold up to international scrutiny.

BehavioralCognitionFieldLongitudinal

C · Applied human research

End-to-end studies for partners

End-to-end applied research for academic and qualified industry partners. Authorship by contribution; data and methods returned to the field.

PartneredPre-registeredIRBEnd-to-end

D · Open work

Protocols, datasets & pipelines

Pre-registered protocols, anonymised datasets, and analysis pipelines — released openly so the next study doesn't begin from zero.

Open sciencePre-registeredReplicablePeer-reviewed

§ 03 · How we work

Slowly, on purpose. Most of the precision is in the watching.

Avg. study
9–18 months

Ethics review
on every project

01 · Frame

Pre-register the question

Define the construct, the population, the resolution required, ethical approvals, and the analytic plan — before a single participant is enrolled. Hypotheses, exclusion rules, and primary outcomes locked in writing.

2–4 wks

02 · Capture

Down to the atom

Validated protocols, calibrated instruments, and pilot cohorts — adapted to the realities of fieldwork in Pakistan. Informed consent in the participant's language. Sensor and survey stacks fused across modalities and audited end-to-end.

3–6 mo

03 · Read

Signal in context

Mixed-effects models against the locked plan, with transparent code reviewed by the people who collected the data. Sensitivity checks, robustness across sites, and full uncertainty reported alongside point estimates.

2–3 mo

04 · Contribute

Back to the record

Plain-language reports for clinical and partner teams; full methods, anonymised datasets, and analysis code for the global research community. Pre-prints first, peer review second, replication encouraged.

1 mo

§ 04 · Contact

If something here resembles a question you're carrying — write to us.

Reach the lab · Direct line

hello@atomresearch.org Copy

Collaborate

Hospitals, academic institutions, and qualified research partners — domestic or international — interested in joint studies. Write with a one-page brief and we'll set up a call to scope feasibility, ethics, and authorship.

Press

For media or peer-review enquiries, please write to the same address with the publication, deadline, and any questions you'd like answered on background.

Open work

Methods, anonymised datasets, and replication code released quarterly under permissive licenses. Subscribe via email for new releases.

Putting Pakistan in the scientific record.