Application of the propensity score in a covariate-based linkage analysis of the Collaborative Study on the Genetics of Alcoholism

Table 1 Overall p-values and the number of significant microsatellite markers by analysis method

Method No.	No. cov.	Covariates analyzed	Overall p-value^b	No. markers with p-values:
				<0.05	<0.01
Single covariate
1	0	none	0.001	16	2
2	1	age_int	0.002	18	3
3	1	sex	0.838	9	1
4	1	maxdrinks	0.002	27	4
5	1	smoker	0.008	19	2
6	1	ttth1	0.055	10	3
Propensity scores^a
7	1	PS1 (age_int, sex)	0.002	22	3
8	1	PS2 (age_int, sex, ttth1)	0.046	12	0
9	1	PS3 (age_int, sex, smoker)	<0.001	27	4
10	1	PS4 (age_int, sex, smoker, maxdrinks)	0.008	21	3
11	1	PS5 (age_int, sex, smoker, maxdrinks, ttth1)	0.329	14	4
Multiple covariates
12	2	age_int, sex	0.025	19	4
13	3	age_int, sex, ttth1	1.000	16	2
14	3	age_int, sex, smoker	0.003	18	4
15	4	age_int, sex, smoker, maxdrinks	0.045	16	4
16	5	age_int, sex, smoker, maxdrinks, ttth1	1.000	9	1

Methods were identified according to the set of covariates analyzed.
^a PS were calculated from the covariates listed, and the corresponding regression coefficients are listed as odds ratios (OR) with the standard errors (SE) in Table 2.
^bThe overall p-value for the analysis method was calculated as the probability of obtaining the observed sum of the LOD scores across the genome in the 1,000 permuted replicates. Significance thresholds of <0.05 and <0.01 were used.

ISSN: 2730-6844