Table 3 Prediction accuracy for specific disorders in patients with at least one diagnosis based on non-coding variants.

From: Application of deep learning algorithm on whole genome sequencing data uncovers structural variants associated with multiple mental disorders in African American patients

 

SNVs in UTR region

SNVs in ncRNA

 

Intronic SNVs

 

Intergenic SNVs

 

Disorder

Precision

Recall

Precision

Recall

Precision

Recall

Precision

Recall

ADHD

38.4 ± 2.5%

33.4 ± 7.3%

39.6 ± 2.7%

32.4 ± 5.7%

39.3 ± 3.6%

31.1 ± 7.1%

42.6 ± 2.3%

42.8 ± 3.9%

Speech/language disorders

30.8 ± 3.3%

31.6 ± 5.8%

30.1 ± 2.9%

26.5 ± 5.3%

30.6 ± 2.8%

31.8 ± 5.4%

33.5 ± 2.8%

33.6 ± 4.5%

Developmental delays

35.4 ± 3.5%

33.5 ± 7.3%

35.4 ± 2.5%

38.4 ± 4.5%

36.6 ± 3.1%

35.5 ± 4.5%

36.4 ± 2.4%

36.8 ± 4.2%

Depression

24.1 ± 2.9%

25.9 ± 7.6%

24.6 ± 2.8%

26.9 ± 4.7%

24.7 ± 3.9%

24.7 ± 5.5%

26 ± 3.1%

26.7 ± 5.0%

Anxiety

18.6 ± 3.7%

17.2 ± 4.2%

17.2 ± 6.6%

6.8 ± 3.9%

16.1 ± 12.3%

3.8 ± 3.2%

20.1 ± 4.4%

17.9 ± 4.4%

ODD

10.9 ± 5.5%

10.3 ± 6.2%

11.6 ± 4.8%

9.6 ± 5.5%

6.8 ± 9.1%

3.2 ± 2.5%

15.4 ± 4.4%

13.5 ± 4.1%

Autism

12.9 ± 4.9%

10.6 ± 5.2%

13.2 ± 4.8%

9.5 ± 3.6%

14.9 ± 7.2%

7.9 ± 4.2%

16.9 ± 5.3%

13.6 ± 4.3%

Intellectual disabilities

19.9 ± 3.4%

14.8 ± 5.3%

22.1 ± 7.9%

10.5 ± 5.3%

21.8 ± 5.8%

9.8 ± 3.5%

20.5 ± 2.9%

22.2 ± 4.4%