Combining gene expression, demographic and clinical data in modeling disease: a case study of bipolar disorder and schizophrenia

Table 1 Demographic and clinical features

Feature	Value (encoding)	Control	Schiz.	Bipolar
Age		44 ± 8	43 ± 9	45 ± 10
Sex	Male (1)	81	86	53
	Female (-1)	31	29	52
PMI		29 ± 13	31 ± 15	38 ± 17
Brain pH		6.6 ± 0.3	6.4 ± 0.3	6.4 ± 0.3
Left brain	Frozen (1)	51	57	59
	Fixed (-1)	61	58	46
Brain region	FrontalBA46 (1)	101	104	94
	FrontalBA46/10 (-1)	11	11	11
HSV 1 OD Z-score		0.1 ± 1.0	-0.2 ± 0.9	-0.0 ± 0.8
HSV 2 OD Z-score		-0.2 ± 0.5	-0.1 ± 0.7	0.3 ± 1.3
Smoking at TOD	Yes (1)	29	71	47
	No (-1)	29	19	18
	Unknown (0)	54	25	40
Alcohol use	Unknown (1)	0	0	4
	Little or none (2)	56	35	12
	Social (3)	38	22	24
	Moderate in past (4)	4	10	16
	Moderate in present (5)	8	10	10
	Heavy in past (6)	6	11	16
	Heavy in present (7)	0	27	23
Drug use	Unknown (1)	0	6	0
	Little or none (2)	97	52	32
	Social (3)	7	7	8
	Moderate in past (4)	5	13	21
	Moderate in present (5)	3	8	12
	Heavy in past (6)	0	11	6
	Heavy in present (7)	0	18	26
Rate of death	Sudden (1)	110	91	96
	Possible anoxia (2)	0	18	6
	Slow death (3)	2	3	0
	Mechanical ventilator (4)	0	3	3

This table lists the distribution (count or mean ± standard deviation) of the demographic features in the three classes (control, schizophrenia, bipolar disorder). Because most classification techniques are restricted to numerical features only, we reencode each nominal feature as a numeric feature. The numerical encoding is listed between parentheses after each nominal feature value.

ISSN: 1471-2164