Table 1 Overview of the statistical analysis on the presence of a facial gestalt for the four ID syndromes using the three described models (one-sided Mann–Whitney U test)

From: Next-generation phenotyping using computer vision algorithms in rare genomic neurodevelopmental disorders

Data set

Model

CIF

Median CIF simulations

U statistic

p value

Koolen–de Vries syndrome

CFPS

0.993

0.8026

389

4.26 × 10−2a

OpenFace

1.143

0.8026

1

2 × 10−4a

Hybrid

1.191

0.8026

0

<1 × 10−4a

PACS1

CFPS

1.908

0.9656

172

1.90 × 10−2a

OpenFace

1.955

0.9538

68

7.6 × 10−3a

Hybrid

2.523

0.9538

5

8 × 10−4a

PPM1D

CFPS

1.011

0.9788

4683.5

4.77 × 10−1

OpenFace

1.667

0.9635

442

4.80 × 10−2a

Hybrid

1.713

0.9635

422.5

4.65 × 10−2a

PHIP

CFPS

2.399

0.9736

59.5

7.1 × 10−3a

OpenFace

1.101

0.9597

3098

3.20 × 10−1

Hybrid

2.239

0.9597

54

6.3 × 10−3a

  1. CIFclustering improvement factor, CFPS Clinical Face Phenotype Space, ID intellectual disability.
  2. aStatistically significant difference.