FSU_Data_analysis

Summary of genotyping

    
      SS  SL  LL
  GG   0   0   1
  AG   1  24  30
  AA  67 147  86

0 people are SgSg, 0 people are SgLg, 1 person is LgLg 1 person is SaSg, 24 people are SaLg, 30 people are LaLg 67 people are SaSa, 147 people are SaLa, 86 people are LaLa

Create effect-coded variables for 3 genotype groups with combined 5-HTTLPR and rs25531: Sx/Sx vs. La/La vs. all other genotypes

Check for normality of distribution of predictor variables:

BDI and NLEQ are both positively skewed. Use square root transformation to normalize both variables.

Center predictor variables

genes$bdi_c <- genes$BDI_tot - mean(genes$BDI_tot, na.rm=T)
genes$nleq_c <- genes$NLEQ_tot - mean(genes$NLEQ_tot, na.rm=T)
genes$nleq_sq_c <- genes$nleq_sq - mean(genes$nleq_sq, na.rm=T)
genes$bdi_sq_c <- genes$bdi_sq - mean(genes$bdi_sq, na.rm=T)

GxE interaction models

Model 1: Gene by NLEQ interaction effect on SIR:

mod1 <- lm(SIR_tot ~ SS_eff*nleq_sq_c + LaLa_eff*nleq_sq_c, data=genes)
summary(mod1)

Call:
lm(formula = SIR_tot ~ SS_eff * nleq_sq_c + LaLa_eff * nleq_sq_c, 
    data = genes)

Residuals:
    Min      1Q  Median      3Q     Max 
-21.737  -7.533  -1.397   5.597  54.567 

Coefficients:
                   Estimate Std. Error t value Pr(>|t|)    
(Intercept)         18.0207     0.6533  27.585  < 2e-16 ***
SS_eff              -0.6098     1.0189  -0.598   0.5499    
nleq_sq_c            2.6109     0.3412   7.651 1.99e-13 ***
LaLa_eff             0.3022     0.9482   0.319   0.7501    
SS_eff:nleq_sq_c    -1.2184     0.5351  -2.277   0.0234 *  
nleq_sq_c:LaLa_eff   1.1778     0.4954   2.377   0.0180 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 10.86 on 347 degrees of freedom
  (6 observations deleted due to missingness)
Multiple R-squared:  0.2056,    Adjusted R-squared:  0.1942 
F-statistic: 17.96 on 5 and 347 DF,  p-value: 7.591e-16

Model 2: Add depression as covariate

mod2 <- lm(SIR_tot ~ SS_eff*nleq_sq_c + LaLa_eff*nleq_sq_c + bdi_sq_c, data=genes)
summary(mod2)

Call:
lm(formula = SIR_tot ~ SS_eff * nleq_sq_c + LaLa_eff * nleq_sq_c + 
    bdi_sq_c, data = genes)

Residuals:
    Min      1Q  Median      3Q     Max 
-21.679  -7.333  -1.311   5.682  53.240 

Coefficients:
                   Estimate Std. Error t value Pr(>|t|)    
(Intercept)         18.0892     0.6450  28.046  < 2e-16 ***
SS_eff              -0.4167     1.0072  -0.414  0.67931    
nleq_sq_c            2.0068     0.3855   5.206 3.32e-07 ***
LaLa_eff             0.2399     0.9359   0.256  0.79782    
bdi_sq_c             1.7632     0.5479   3.218  0.00141 ** 
SS_eff:nleq_sq_c    -1.2164     0.5281  -2.304  0.02184 *  
nleq_sq_c:LaLa_eff   1.2653     0.4897   2.584  0.01017 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 10.72 on 346 degrees of freedom
  (6 observations deleted due to missingness)
Multiple R-squared:  0.2287,    Adjusted R-squared:  0.2153 
F-statistic:  17.1 on 6 and 346 DF,  p-value: < 2.2e-16

Model results hold when depression entered as covariate.

Compare models 1 and 2

	Dependent variable:
	SIR_tot
	(1)	(2)
Constant	18.021***	18.089***
	(0.653)	(0.645)
SS	-0.610	-0.417
	(1.019)	(1.007)
NLEQ	2.611***	2.007***
	(0.341)	(0.386)
LaLa	0.302	0.240
	(0.948)	(0.936)
BDI		1.763***
		(0.548)
SS*NLEQ	-1.218**	-1.216**
	(0.535)	(0.528)
LaLa*NLEQ	1.178**	1.265**
	(0.495)	(0.490)
Observations	353	353
R2	0.206	0.229
Adjusted R2	0.194	0.215
Residual Std. Error	10.859 (df = 347)	10.716 (df = 346)
F Statistic	17.962*** (df = 5; 347)	17.098*** (df = 6; 346)
Note:	p<0.1; p<0.05; p<0.01

Interaction plots

Scale for 'colour' is already present. Adding another scale for
'colour', which will replace the existing scale.

LaLa group displays a greater susceptibility to the impact of life stress (NLEQ) on self-control, whereas SS group less reactive to life stress in terms of effects on hoarding.

Specificity quesition

Rerun models 1 and 2 substituting OCIR as outcome variable

mod4 <- lm(OCIR_tot ~ LaLa_eff*nleq_sq_c + SS_eff*nleq_sq_c, data=genes)
summary(mod4)

Call:
lm(formula = OCIR_tot ~ LaLa_eff * nleq_sq_c + SS_eff * nleq_sq_c, 
    data = genes)

Residuals:
    Min      1Q  Median      3Q     Max 
-16.252  -6.300  -2.305   4.780  30.349 

Coefficients:
                   Estimate Std. Error t value Pr(>|t|)    
(Intercept)         12.8020     0.5192  24.656  < 2e-16 ***
LaLa_eff             0.1231     0.7537   0.163    0.870    
nleq_sq_c            1.4650     0.2712   5.402 1.23e-07 ***
SS_eff              -0.4613     0.8098  -0.570    0.569    
LaLa_eff:nleq_sq_c  -0.1674     0.3938  -0.425    0.671    
nleq_sq_c:SS_eff     0.1743     0.4253   0.410    0.682    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 8.631 on 347 degrees of freedom
  (6 observations deleted due to missingness)
Multiple R-squared:  0.1013,    Adjusted R-squared:  0.08835 
F-statistic: 7.823 on 5 and 347 DF,  p-value: 5.409e-07

mod5 <- lm(OCIR_tot ~ LaLa_eff*nleq_sq_c + SS_eff*nleq_sq_c + bdi_sq_c, data=genes)
summary(mod5)

Call:
lm(formula = OCIR_tot ~ LaLa_eff * nleq_sq_c + SS_eff * nleq_sq_c + 
    bdi_sq_c, data = genes)

Residuals:
    Min      1Q  Median      3Q     Max 
-15.604  -5.769  -1.907   3.812  27.992 

Coefficients:
                   Estimate Std. Error t value Pr(>|t|)    
(Intercept)        12.87531    0.50638  25.426  < 2e-16 ***
LaLa_eff            0.05647    0.73477   0.077  0.93879    
nleq_sq_c           0.81906    0.30267   2.706  0.00714 ** 
SS_eff             -0.25489    0.79075  -0.322  0.74739    
bdi_sq_c            1.88548    0.43013   4.384 1.55e-05 ***
LaLa_eff:nleq_sq_c -0.07385    0.38443  -0.192  0.84778    
nleq_sq_c:SS_eff    0.17644    0.41458   0.426  0.67068    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 8.413 on 346 degrees of freedom
  (6 observations deleted due to missingness)
Multiple R-squared:  0.1486,    Adjusted R-squared:  0.1338 
F-statistic: 10.06 on 6 and 346 DF,  p-value: 2.948e-10

Nonsignificant results in OCIR models suggesting that GxE effects seem to be specific to hoarding.