Schizophrenia Candidate Gene Regions: Methods and Resources

Datamining and Database Development

There have been 47 independent genomewide scans published for linkage to schizophrenia (SZ) from 1994 to 2010. In addition, there have been three meta-analyses of SZ genome scan completed, and at least 27 microarray studies conducted to test for differential expression of genes between schizophrenics and unaffected individuals, and more than 1700 association studies on candidate SZ genes initially identified based on their genomic position or biological functions. We carefully identified and reviewed available literature. The database, which was created by using the results of our data analyses and simulations, was incorporated into this website.

Determining the Schizophrenia Candidate Regions (SCRs)

Data mined from the 47 genome scans included genetic markers that showed either ‘significant’ or ‘suggestive’ genome-wide evidence of linkage to schizophrenia based on the significance threshold established by Lander & Kruglyak (1995). We then projected these candidate markers onto the Rutgers Map. We utilized three methods to define the SCRs: single significant hit approach, disjoint approach, and smoothing method. Through these methods, we identified 22 SCRs.

In-House Resources

For general computing, we utilized one of the Computational Genetics servers ( that employs a quad-core 2.66GHz Intel Xeon processor X3330 PC with 6GB RAM. For the simulations, we used the departmental cluster ( that employs multi-core (quad-core, 8-core, 12-core, and 16-core) processor PCs amounting to a total of 262 computational nodes (total memory: 845.6 GB RAM; total disk: 4755.1 GB) running the Ubuntu 10.04.1 LTS (lucid) operating system (OS).

We also used the Rutgers High-Resolution Combined Linkage-Physical Map and MAP-O-MAT. The existing Rutgers Map (Rutgers Map v.2), released in 2006 (N=28219, NCBI Build 36, hg18, March 2006), is the largest combined linkage-physical map of the human genome (Kong et al. 2004; Chen et al. 2006; Matise et al. 2007) and it has been been incorporated into MAP-O-MAT automated linkage mapping server (Kong & Matise 2005). After updating the Rutgers Map (N=27991, NCBI Build 37, hg19, Feb 2009/Aug 2010), we utilized it as a backbone for localizing the results of the studies used to determine and characterize our SCRs.

© 2014 Matise Laboratory of Computational Genetics