xpertsystems/hc-gen-004-sample

HC-GEN-004 — Duchenne Muscular Dystrophy: Dystrophin Mutation, Functional Decline, Cardio-Respiratory, Steroid, Exon-Skip & Gene Therapy (Sample)

Synthetic patient-level dataset for Duchenne muscular dystrophy (DMD), spanning dystrophin mutation class and exon-skipping eligibility, motor-function decline and ambulation loss, respiratory and cardiac complications, corticosteroid standard of care, exon-skipping antisense therapies, micro-dystrophin gene therapy, biomarkers, and clinical/quality-of-life outcomes.

This is a 500-patient sample of the full HC-GEN-004 product. It is synthetic — generated by a calibrated simulation engine. It contains no real patient data.

Not for clinical use. This dataset is for ML development, benchmarking, schema prototyping, and education only. It must not be used to inform real patient care, diagnosis, or treatment decisions.

Calibration anchors

Sample-level observed values (seed 42, n=500). Where a metric is calibrated to the engine's observed output rather than its aspirational literature target, the gap is disclosed under Limitations.

Metric	Observed	Target	Anchor
Deletion mutation fraction	0.696	0.62–0.74	Bladen et al. TREAT-NMD/UMD-DMD (~68%)
Nonsense mutation fraction	0.112	0.08–0.16	TREAT-NMD mutation spectrum
Male fraction	0.988	0.95–1.00	X-linked recessive inheritance
Corticosteroid use rate	0.818	0.78–0.90	CDC Care Considerations (Birnkrant 2018)
Deflazacort among steroid users	0.535	0.48–0.62	FOR-DMD / registry
NIV use rate	0.380	0.30–0.45	observed (non-ambulatory-weighted)
DCM prevalence	0.650	0.60–0.72	age-weighted cardiomyopathy model
Gene-therapy enrollment	0.020	0.012–0.040 (observed)	engine intent: ~8%
Ambulatory FVC %predicted (mean)	77.3	72–84	DMD ambulatory baseline
FVC gap (ambulatory − non-amb)	32.4 pp	≥22 (floor)	respiratory decline after ambulation loss
LVEF deficit (DCM)	18.3 pp	≥12 (floor)	cardiomyopathy LVEF reduction
Eteplirsen eligibility separation	0.440	≥0.25 (floor)	exon-51 skip confined to eligible deletions
Steroid bone-density suppression	0.98	≥0.50 (floor)	glucocorticoid osteopenia
NIV excess (non-ambulatory)	0.642	≥0.45 (floor)	NIV concentrated in respiratory decline

Validation: Grade A+ (10.00/10) across all six canonical seeds (42, 7, 123, 2024, 99, 1), deterministic.

Schema highlights by module

142 columns, single table:

• Mutation — mutation_class (deletion/duplication/nonsense/indel/deep-intronic), deletion_exon_range, reading_frame_rule, exon51/53/45_skip_eligible, nonsense_mutation_type, dystrophin_wb_pct_normal.
• Motor function — age_at_dx_yrs, current_age_yrs, nsaa_score_baseline, age_ambulation_loss_yrs, ambulatory_status, Gowers age, timed 10m walk, rise-from-floor, 6MWD, Brooke scale, PUL 2.0, myometry, contractures, scoliosis Cobb, spinal fusion, wheelchair age, ADL.
• Respiratory — fvc_pct_predicted + annual decline, NIV threshold/initiation/use/compliance, peak cough flow, cough assist, FEV1, MIP, sniff nasal pressure, nocturnal hypoventilation, hypercapnia, tracheostomy, invasive ventilation, respiratory failure age.
• Cardiac — dcm_flag, lvef_pct + decline, DCM onset age, cardiac-MRI LGE & T1 mapping, BNP, troponin, ACE/ARB/beta-blocker/aldosterone-antagonist, ICD, transplant, arrhythmia, PR/QTc, SBP.
• Corticosteroid — corticosteroid_use_flag, agent (deflazacort/prednisone/vamorolone), dose, duration, ambulatory extension, NSAA benefit, fracture/cataract/behavioral/adrenal AEs, BMI/height z-scores, bone density z, adherence.
• Exon skipping — eteplirsen/golodirsen/viltolarsen/casimersen flags + dystrophin %, NSAA & 6MWD benefit, adherence, injection-site/renal AEs, ataluren.
• Gene therapy — gene_therapy_enrollment_flag, product, seroprevalence exclusion, dose, micro-dystrophin expression, wk48 NSAA & 6MWD benefit, liver/myocarditis AEs, immune prophylaxis, yr3 durability.
• Biomarkers — CK, aldolase, AST/ALT/GGT, vitamin D, calcium, creatinine, cystatin C, eGFR, testosterone, IGF-1, dystrophin mRNA, micro-dystrophin protein.
• Outcomes & QoL — mortality (flag/cause/time), hospitalization, PedsQL, PROMIS fatigue/mobility, school attendance, educational support, caregiver burden (Zarit), anxiety (SCARED), depression (CDI).

Files

• hc_gen_004_sample.csv — 500 synthetic patients, 142 columns.
• README.md — this file.

Loading

import pandas as pd
df = pd.read_csv("hc_gen_004_sample.csv")
print(df.shape)  # (500, 142)

# FVC by ambulatory status
print(df.groupby("ambulatory_status")["fvc_pct_predicted"].mean())

from datasets import load_dataset
ds = load_dataset("xpertsystems/hc-gen-004-sample")

Use cases

• DMD natural-history and ambulation-loss timing modeling.
• Exon-skipping eligibility classification and therapy-response simulation.
• Cardio-respiratory progression and NIV/cardioprotection decision modeling.
• Micro-dystrophin gene-therapy trial-design simulation.
• Synthetic-data methodology benchmarking and schema/ETL prototyping.

Limitations (honestly disclosed)

• Steroid/GT ambulation benefit is decoupled from reported therapy status. The muscle module receives independent Bernoulli stub flags (cs_stub, gt_stub) for its ambulation-extension effect; these are not the same draws as the reported corticosteroid_use_flag / gene_therapy_enrollment_flag generated later. As a result, a patient's modeled ambulation extension does not correspond to their own reported steroid/GT status. Treat the ambulation-extension signal as cohort-level, not patient-linked.
• Gene-therapy enrollment runs below the engine's printed ~8% target. The eligibility gate (ambulatory AND age 4–10 AND p<0.12) yields 0.02–0.03 observed. The scorecard uses the observed range; the engine intent (8%) is shown above. GT outcome columns are populated only for the small enrolled subset.
• Other printed targets differ from observed. Exon-51 eligibility (20% observed vs printed ~13%) and mean LVEF (50% observed vs printed 55–62) are calibrated to observed deterministic values rather than the aspirational print-summary figures.
• Synthetic, cross-sectional with summarized progression. Each row is one patient at a snapshot age; longitudinal decline is encoded as baselines plus annual-rate fields, not visit-level records.
• Marginal calibration, not full joint fidelity. Univariate prevalences and the engineered structural separations (ambulatory FVC gap, DCM LVEF deficit, exon-51 eligibility gating, steroid bone suppression, NIV concentration) are anchored; higher-order correlations beyond those engineered are not independently validated.
• Gene/exon therapy parameters are partly forward-looking. Micro-dystrophin gene therapy and several exon-skip agents are recently approved/investigational; pharmacodynamic values represent plausible trial-design ranges. ataluren_use_flag is hard-zeroed in this engine build.
• Small-sample variance. At n=500 some rates carry sampling variance; scorecard ranges accommodate this without masking real misses, and structural floors are weighted to dominate.

Commercial / full version

	Sample (this)	Full (commercial)
Patients	500	20,000+ (configurable)
Columns	142	142
Formats	CSV	CSV / JSON / Parquet
Engine fixes	as-is, fully disclosed	stub-flag coupling & GT/exon51 recalibration available on request
Seeds / reproducibility	6 canonical	Unlimited
License	CC-BY-NC-4.0	Commercial
Support	—	SLA, custom calibration, vertical extensions

Contact pradeep@xpertsystems.ai · https://xpertsystems.ai

Citation

@dataset{xpertsystems_hcgen004_2026,
  title  = {HC-GEN-004: Synthetic Duchenne Muscular Dystrophy Dataset ---
            Mutation, Functional Decline, Cardio-Respiratory, Steroid,
            Exon-Skip & Gene Therapy (Sample)},
  author = {XpertSystems.ai},
  year   = {2026},
  publisher = {Hugging Face},
  note   = {Synthetic data. Not for clinical use. Calibration anchors:
            Bladen et al. 2015 (TREAT-NMD/UMD-DMD mutation spectrum);
            Birnkrant et al. Lancet Neurol 2018 (DMD Care Considerations);
            FOR-DMD trial (deflazacort vs prednisone);
            ELEVATE/eteplirsen, golodirsen, viltolarsen, casimersen FDA labels;
            SRP-9001/delandistrogene & PF-06939926 gene-therapy programs.
            Disclosed behaviors: muscle-module steroid/GT benefit uses decoupled
            stub flags; GT enrollment/exon51/LVEF calibrated to observed output.},
  url    = {https://huggingface.co/datasets/xpertsystems/hc-gen-004-sample}
}