WhelWomen's Health Evidence Lab

Whel is building a drug-repurposing platform for female biology that surfaces approved drugs already working for women's health conditions, validates them against mechanistic and clinical evidence, and maps each to the 505(b)(2) regulatory pathway.

Read the manifesto →See the candidates →
111
signals indexed · 6 conditions covered
433
verbatim claims, each pinned to a source quote
80%
of preclinical research uses male-only animal models
1993
women excluded from most US Phase I & II trials through this year
The thesis

Drug development was built on male biology.

Every AI drug-discovery platform reasons over a knowledge graph built from male-default research, and rather than adding a women's-health filter to someone else's substrate, we are building the corrected version from the ground up, grounded in female biology.

Inherited
Male-default knowledge graph, sparse for female biology.
Corrected
Sex-specific drugs, mechanisms, and the relations between them.
How female biology was written out of drug development →

The drugs that work for women's health were never developed for it.

01
Metformin → PCOS
insulin resistance, anovulation
02
Spironolactone → Hormonal acne
anti-androgen, off-label staple
03
GnRH antagonists → Endometriosis
from prostate cancer to standard of care
04
SSRIs → PMDD
neurosteroid modulation, not reuptake
05
Letrozole → PCOS infertility
from breast cancer; superior to clomiphene
06
GLP-1 agonists → PCOS · Endometriosis
metabolic + inflammatory, under study
Why a historical precedent led us to start with drug repurposing →
The output

What the platform has surfaced.

See all candidates →

Two candidates, chosen to show the range the score captures: one of the strongest signals on the platform, and one rated lower, where the evidence is thinner or the independent readings disagree. The score separates them. Open either for the full evidence trail; the full index has the rest.

WHEL-C-086
Trials · Phase 3InvestigationalClinically anchoredStrong tier
D-chiro-inositol→PCOS
Origin · Existing drug · repurposing candidate
Scored for women
Evidence generated in women · (female population, ~100% female).
×1.00
F1 multiplier

A systematic review/meta-analysis and a comparative treatment study indicate D-chiro-inositol provides benefits for some metabolic measures, ovulation, menstrual regularity, and insulin resistance in PCOS patients.

Signal type◂ anchors the headline
Direct ◂9.0 · Str
Composite9/10Evidence · 4 verbatim claims
Score by metric · Direct arm
Corroboration2/2
Rigor2/2
Specificity2/2
Plausibility1/2
Consistency2/2
Clinical-trial status · for this condition

Studied as a therapy for pcos in 4 interventional trials · highest reached Phase 3 · completed.

Source: ClinicalTrials.gov (U.S. National Library of Medicine).

Regulatory & development status

No FDA-approved label. No FDA-approved drug label was found for this molecule (it may be a supplement, a biologic, investigational, or approved only outside the US), so any use here is investigational.

Not in the Orange Book. Absent from the Orange Book. Biologics (listed in the Purple Book instead), dietary supplements, and drugs not FDA-approved in the US do not appear here.

This is where the candidate sits in the regulatory landscape: descriptive context, not a 505(b)(2) viability assessment or regulatory advice.

Sources: FDA Orange Book; DailyMed (U.S. National Library of Medicine).

Full breakdown →
WHEL-C-003
Off-label · genericUnvalidated signalModerate tier
ULIPRISTAL ACETATE→PMDD
Origin · Approved
Scored for women
Evidence generated in women · (female population).
×1.00
F1 multiplier

Ulipristal acetate is a Phase 2 clinical candidate for PMDD acting as a progesterone receptor modulator on the progesterone receptor per Open Targets.

Signal type◂ anchors the headline
Pathway ◂6.0 · Mod
Composite6/10Evidence · 1 verbatim claim
Score by metric · Pathway arm
Corroboration0/2
Rigor1/2
Specificity2/2
Plausibility2/2
Consistency1/2
Regulatory & development status

Off-label. The drug has an FDA-approved label, but for a different indication, so using it for this condition would be off-label.

Generic available. A live ANDA lists this active ingredient, so the basic molecule is available as a generic (off-patent in its basic form).

This is where the candidate sits in the regulatory landscape: descriptive context, not a 505(b)(2) viability assessment or regulatory advice.

Sources: FDA Orange Book; DailyMed (U.S. National Library of Medicine).

Full breakdown →
How it works

Built in three layers.

Layer 01
The substrate

A corrected knowledge graph built to hold sex-specific pharmacokinetics, cyclical hormonal state, and the cross-condition mechanistic relationships general platforms miss because they were trained on male-default data. Grounded today in MONDO, EFO, RxNorm, and ChEMBL, with sex-specific pharmacokinetics and cyclical-phase layers already seeded and shown beside the relevant signals, and the fuller female-specific structure no existing ontology covers still being built in.

Sources · Open Targets, FDA drug labels, the curated sex-PK literature
Postgres-nativeOntology-groundedSex-PK + cycle-phase
Layer 02
Retrieval & validation

Provenance-preserving extraction tuned for biomedical literature. Every claim ties to a verbatim source span, every synthesis is marked as a synthesis, and every contradiction in the underlying literature is surfaced explicitly rather than averaged. This is what the §3060 research-support exemption requires and what clinicians need to trust the output.

Sources · PubMed, ClinicalTrials.gov, FDA openFDA
Per-claim provenanceMarked synthesisContradiction surfacing
Layer 03
Hypothesis from signal

Patient-community signal, including off-label prescribing patterns, community reports, and structured patient-reported data, enters as hypothesis generation and is validated downstream against mechanistic and clinical evidence. It is never equated with the results of a controlled trial, and it is the input that surfaces the hypotheses worth checking. Formal advocacy-organization partnerships are planned, taken on once the validation work is in place.

Sources · Patient-community reports (Reddit)
Off-label patternsCommunity reportsValidated downstream
Beside every signal
Independent validation

Once a signal is built by the three layers above, it is checked against outside references that are kept separate from its score: an external validation ladder (E0–E3) that traces to guideline bodies such as ESHRE, ACOG, and Cochrane, and Every Cure’s MATRIX treatment-probability model. This is the validation layer. It is shown beside each result, kept out of the score, and is not one of the three build layers.

References · Every Cure MATRIX, ESHRE / ACOG / Cochrane guidance
External validation ladder · E0–E3MATRIX cross-referenceDRKG · PrimeKG · TxGNN planned

From off-label signal to a 505(b)(2)-eligible candidate.

Off-label use is the largest uncontrolled clinical trial in women's health. We read the results.

01→
Signal
Off-label & patient signal
Prescribing patterns, community reports, under-read literature.
02→
Ground
Ontology grounding
Entities resolved to MONDO, EFO, RxNorm, ChEMBL.
03→
Validate
Mechanistic & clinical check
Per-claim provenance; synthesis marked; contradictions surfaced.
04→
Tier
Confidence tiering
Graded from strong to exploratory, never flattened into one weight.
05
Candidate
505(b)(2)-eligible candidate
Which drug, which condition, with its full evidence trail.
v0 corpus · 6 conditions

Where we start.

All conditions →
C-0111 signals
Adenomyosis
Endometrial tissue invades the uterine muscle wall, causing an enlarged uterus, severe period pain, heavy bleeding, and chronic pelvic pain. It frequently coexists with endometriosis and is a major cause of iron deficiency anemia.
11 signalsOpen →
C-0218 signals
Endometriosis
Endometriosis is a chronic inflammatory disease in which tissue similar to the uterine lining grows outside the uterus, on the ovaries, bowel, bladder, and peritoneum. These lesions bleed with every menstrual cycle, causing progressive scarring, adhesions, and infertility.
18 signalsOpen →
C-0321 signals
PCOS
PCOS is the most common endocrine disorder in women of reproductive age. It involves elevated androgens, irregular or absent ovulation, and polycystic ovaries. Presentations range from lean women with menstrual irregularity to those with severe insulin resistance and metabolic syndrome. It is the leading cause of anovulatory infertility worldwide.
21 signalsOpen →
C-0437 signals
Perimenopause & Menopause
Perimenopause is the 4 to 10 year transition before menopause, during which hormones fluctuate and symptoms begin. Menopause is confirmed after 12 consecutive months without a period. Hot flashes and night sweats are the hallmark complaints. Genitourinary changes, mood disruption, and insomnia are common and frequently undertreated.
37 signalsOpen →
C-05Flagship
PMDD
PMDD is a severe cyclic disorder in which the brain responds abnormally to normal hormonal changes in the luteal phase. Symptoms include depression, anxiety, and irritability that appear reliably before menstruation and resolve within days of onset. It is distinct from PMS in severity and carries elevated suicide risk in the luteal phase.
10 signalsOpen →
C-0614 signals
Vulvodynia
Vulvodynia is chronic vulvar pain lasting at least three months with no identifiable cause. The most common form is provoked vestibulodynia, where pain is triggered by touch at the vaginal opening. It is a leading cause of sexual dysfunction and gynecologic care avoidance.
14 signalsOpen →
Fig. 1 · Confidence distribution

Where the evidence sits.

Evidence sits in tiers, and we never flatten strong and exploratory signals into the same visual weight.

Color intensity scales with signal count within each tierClick a row to open the condition
The expansion path

Repurposing is where we prove the platform.

Phase 1
Drug repurposing for core women's health
PMDD, endometriosis, PCOS, adenomyosis, perimenopausal mood disorders, vulvodynia. Output: specific candidates with full evidence trails for the researchers and women's health teams working on these conditions.
Phase 2
Expand to sex-divergent conditions
Autoimmune disease (80% female), pain conditions (1.5–4× female prevalence), neuropsychiatric conditions. The knowledge graph and sex-specific PK modeling transfer directly, the same platform extended to larger patient populations.
Phase 3
Beyond repurposing
Novel target identification, combination-therapy prediction, and sex-stratified clinical-trial design, all capabilities latent in the substrate built for repurposing. The operating system for female-biology drug development.
See the full roadmap →
Whel · Women's Health Evidence Lab

Finding what already works for women.

For clinician-researchers, pharma women's health teams, and advocacy organizations. Explore the full candidate index, or read how each signal is scored.

View candidates →How we score evidence