Performance and clinical utility of a new supervised machine-learning pipeline in detecting rare ciliopathy patients based on deep phenotyping from electronic health records and semantic similarity

Table 2 Top 20 HPO concepts among cases and controls

Ciliopathy cases (n = 30)		Controls (n = 7231)
HPO concepts	# Patients (%)	HPO concepts	# Patients (%)
Chronic kidney disease	77	Fever	47
Abnormality of the kidney	73	Pyelonephritis	32
Nephropathy	73	Proteinuria	31
Renal insufficiency	73	Pain	26
Anemia	53	Seizure	17
Fatigue	53	Renal insufficiency	15
Proteinuria	53	Vesicoureteral reflux	15
Arteriovenous fistula	47	Muscular hypotonia	15
Stage 5 chronic kidney disease	43	Cough	14
Hypertension	40	Diarrhea	14
Hyperparathyroidism	37	Hematuria	13
Pain	37	Constipation	12
Pollakisuria	37	Abnormality of the kidney	12
Anorexia	30	Increased body weight	12
Fever	30	Hypertension	12
Nausea	30	Moderate albuminuria	11
Polydipsia	30	Edema	11
Enuresis	27	Albuminuria	11
Hypokalemia	27	Nephropathy	11
Moderate albuminuria	27	Respiratory distress	11

ISSN: 1750-1172