**
Instructions for running models described In:**

**Devlin, B., Bacanu, S., Roeder, K., "Genomic
Control in the Extreme," Nature Genetics, 2004, 36(11):1129-1130.**

**Bacanu, S., Devlin, B., Roeder, K.,
"Association Studies for Quantitative Traits in Structured Populations,"
Genetic Epidemiology, 2002, 22:78-93.**

**Devlin, B. and Roeder, K., "Genomic Control
for Association Studies," Biometrics, 1999, 55:997-1004.**

**Overview**

The GC program is provided below.

GC implements the genomic control models introduced in Devlin and Roeder (1999) and further expanded in Bacanu et al. (2002). The program runs in two steps. First, it estimates the inflation factor(s) from a set of "null" loci specified by the user. The program then uses these estimates to test for effects at either a single locus (marginal) or two loci (interaction) with relation to a user specified response variable.

GC can optionally adjust the p-values for uncertainty in the estimated effect of substructure. This is described in Devlin et al. (2004). (Note: This adjustment was previously achieved by running a second progrm (GCF). However, both programs have been merged, and GCF behavior is achieved by setting an option in GC.) This approach is preferable if a large number of tests are conducted.

This program is implemented in the **R**
programming language (freeware version of S-Plus) for Linux (version 1.7.0).
Thus, the first step in running the program will be to download a copy of R if
you do not already have it installed. To download R, go to

and follow the necessary links to download R.

Once R is installed, simply type R at the command prompt to start (Linux) or select R from the Start Menu (Windows). To quit R,
type **q()** at the R command prompt. To cancel an R command, type *
control-c*.

**Example 1**

Marginal models with covariates, binary response, 100 loci (90 null).

**Model output file
(including summary statistics)**

**Example 2**

Marginal models with covariates, binary response, 100
loci (all null by default).

P-values are adjusted.

This example uses the same data as that in Example 1.

**Model output file
(including summary statistics)**

**Example 3**

Interaction models with covariates, binary response, 50
loci (45 null).

(By default, p-values are not adjusted for uncertainty.)

**Model output file
(including summary statistics)**

**Example 4**

Marginal models with covariates, gaussian response, 100 loci (30 null).

The inflation factor is assumed to be 1. (By default, p-values aren't adjusted.)

**Model output file
(including summary statistics)**

**Example 5**

Marginal models with covariates, gaussian response, 100
loci (all null by default).

P-values are adjusted for uncertainty.

This example uses the same data as that in Example 4.