On the Integration of Alcohol-Related Quantitative Trait Loci and Gene Expression Analyses Robert Hitzemann, Cheryl Reed, Barry Malmanger, Maureen Lawler, Barbara Hitzemann, Brendan Cunningham, Shannon McWeeney, John Belknap, Christina Harrington, Kari Buck, Tamara Phillips, and John Crabbe Background: Quantitative trait loci (QTLs) have been detected for a wide variety of ethanol-related phenotypes, including acute and chronic ethanol withdrawal, acute locomotor activation, and ethanol preference. This study was undertaken to determine whether the process of moving from QTL to quanti- tative trait gene (QTG) could be accelerated by the integration of functional genomics (gene expression) into the analysis strategy. Methods: Six ethanol-related QTLs, all detected in C57BL/6J and DBA/2J intercrosses were entered into the analysis. Each of the QTLs had been confirmed in independent genetic models at least once; the cumulative probabilities for QTL existence ranged from 10 -6 to 10 -15 . Brain gene expression data for the C57BL/6 and DBA/2 strains (n = 6 per strain) and an F 2 intercross sample (n = 56) derived from these strains were obtained by using the Affymetrix U74Av2 and 430A arrays; additional data with the U74Av2 array were available for the extended amygdala, dorsomedial striatum, and hippocampus. Low-level analy- sis was performed by using multiple methods to determine the likelihood that a transcript was truly differentially expressed. For the 430A array data, the F 2 sample was used to determine which of the differentially expressed transcripts within the QTL intervals were cis-regulated and, thus, strong candidates for QTGs. Results: Within the 6 QTL intervals, 39 transcripts (430A array) were identified as being highly likely to be differentially expressed between the C57BL/6 and DBA/2 strains at a false discovery rate of 0.01 or better. Twenty-eight of these transcripts showed significant (logarithm of odds 3.6) to highly significant (logarithm of odds 7) cis-regulation. The process correctly detected Mpdz (chromosome 4) as a candidate QTG for acute withdrawal. Conclusions: Although improvements are needed in the expression databases, the integration of QTL and gene expression analyses seems to have potential as a high-throughput strategy for moving from QTL to QTG. Key Words: QTL, Gene Expression, Ethanol, Mice, Brain. O VER THE PAST 10 years, more than 120 behavioral quantitative trait loci (QTLs) have been detected in various mouse crosses and confirmed at logarithm of odds (LOD) scores of more than 3 (Flint, 2003). This impressive accomplishment has contributed greatly to the now widely held view that the genetic influence on most behavioral phenotypes involves the actions of multiple gene products, each with a moderate to small effect (e.g., Belknap et al., 2001). However, the transition from QTL to quantitative trait gene (QTG) and eventually to quantitative trait nu- cleotide (Mackay, 2001) has been particularly slow and the subject of considerable debate (Nadeau and Frankel, 1999). What may be considered the traditional approach (Darvasi, 1998) involves capturing the QTL in congenic and interval-specific congenic mouse strains, followed by the application of molecular techniques to detect functional relevant polymorphisms within the reduced QTL interval. To our knowledge, this approach has been successfully applied only once in the mouse to a behavioral trait (Fehr et al., 2002; Shirley et al., 2004). The traditional approach has several problems, including failure to capture the QTL within the congenic strains due to changes in genetic back- ground; in addition, producing the congenic and recombi- nant congenic strains can be expected to take several years and is thus vulnerable to experimental attrition. The earliest proof of principle (in rodent models) of an alternative approach, namely, the integration of QTL and expression data, was two studies that identified the genes involved in insulin resistance (Aitman et al., 1999; Collison From the Department of Behavioral Neuroscience and Portland Alcohol Research Center (RH, CR, BM, ML, BH, BC, JB, KB, TP, JC), Department of Public Health and Epidemiology (SM) and Gene Microarray Shared Resource (CH), Oregon Health & Science University, Portland, Oregon; and the Research Service (RH, JB, KB, TP, JC), Veterans Affairs Medical Center, Portland, Oregon. Received for publication November 5, 2003; accepted June 7, 2004. Supported in part by NIAAA Grants DA 05228, AA 11043, AA 11384, AA 10760, AA 06243, and AA 11114 and by the Department of Veterans Affairs. Reprint requests: Robert Hitzemann, PhD, Department of Behavioral Neuroscience, Oregon Health & Science University, Portland, OR 97201; Fax: 503-494-6877; E-mail: hitzeman@ohsu.edu. Copyright © 2004 by the Research Society on Alcoholism. DOI: 10.1097/01.ALC.0000139827.86749.DA 0145-6008/04/2810-1437$03.00/0 ALCOHOLISM:CLINICAL AND EXPERIMENTAL RESEARCH Vol. 28, No. 10 October 2004 Alcohol Clin Exp Res, Vol 28, No 10, 2004: pp 1437–1448 1437