Abraham A. Palmer - US grants

This proposal is designed to address the possible genetic relationships between the locomotor stimulant effects of ethanol and those of allopregnanolone. We have already analyzed the locomotor stimulant effects of these two compounds in the BXD recombinant inbred mouse set, and identified a statistically significant correlation between the locomotor response to ethanol and locomotor response to allopregnanolone. This analysis has identified a number of provisional quantitative train loci for response to allopregnanolone, some of which are regions known to contain loci for ethanol traits. Here we propose to extend upon these, findings by generating an F2 population which will allow us to confirm or reject these provisional QTLs. Additionally, we will examine locomotor response to allopregnanolone in FAST and SLOW mice selected for their high and low locomotor response to ethanol in order to determine whether selection has favored similar responses to allopregnanolone. We will also determine whether cross sensitization exists between ethanol and allopregnanolone. Finally, we will test finasteride's ability to antagonize the locomotor stimulant effects of ethanol.

[unreadable] DESCRIPTION (provided by applicant): [unreadable] The proposed 4-year training award is intended to provide the applicant with additional training in three areas: microarray study design and analysis, bioinformatic techniques, and translational genetics. Experts in all of these fields have agreed to participate in the training program. Additionally, structured course work is proposed that addresses the trainee's deficiencies. The proposed training environment is an interdisciplinary center established by Columbia University to promote interaction among biologists, geneticists, statisticians, computational biologists, bioinformaticians, and engineers for the purpose of forging interdisciplinary "genomic" solutions to biomedical problems. Experts in all relevant fields are already working on closely related projects under the umbrella of a program project grant that has been awarded to the sponsor of this training program. Thus, the environment is ideally suited to the training goals. The applicant is already expert in the tools of classical behavioral genetics, such as quantitative trait locus (QTL), analysis and the creation of selected lines. The research plan is designed to identify genes that influence naturally occurring variability in a genetically tractable form of fear learning in mice. Short term selected mouse lines will be created and used to identify QTL, and gene expression differences. Specific polymorphisms related to genes and gene expression will be identified by the synergistic application of traditional (QTL mapping) and modern (microarray and bioinformatics) techniques. By using the latter techniques, which are the training component of this application, specific genes will be rapidly identified. This proposal addresses the major weakness of traditional mouse QTL studies, which is that they are seldom powerful enough to identify specific genes. We predict that some of the identified genes will also control fear learning and anxiety disorders in human subjects. To test these hypotheses, the strongest candidate genes, gene classes, and pathways will be examined for polymorphisms in a large population (approximately 1000) of unrelated normal human subjects scored for fear learning, and other anxiety dimensions, as well as in a population of anxiety disorder patients. Our collaborators are ascertaining these human subjects as one component of the sponsor's program project grant. This training program will prepare the candidate for an academic career focused on the use of endophenotypes, animal models and bioinformatics to elucidate the genetic basis of psychiatric disease. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): We propose a multidisciplinary, translational genetic study to identify genes that influence sensitivity to the stimulant drugs methamphetamine and amphetamine in mice and humans. We will use mice to better define chromosomal regions (also called quantitative trait loci or QTL) that contain genes that influence sensitivity to the acute locomotor stimulant effects of methamphetamine. This will be accomplished using two complimentary approaches: congenic strains, and advanced intercross lines. Using these techniques, we will be able to determine, with some precision, the location of genes that influence this trait. We will then use a variety of experimental and bioinformatic approaches to parse among the genes located in those regions, which will allow us to identify a small number of high probability candidate genes. We expect that this list of genes will include those responsible for differences in the response to methamphetamine in the mice. We will then explore the effect of polymorphisms in these genes on the subjective response to amphetamine in a sample of carefully screened, healthy human volunteers. All volunteers will give informed consent prior to participation in the study. We will use a double-blind, within-subject design in which each subject receives each dose (placebo, 5,10 and 20 mg) of amphetamine once, in a randomized order, over the course of 4 sessions. During the sessions when drug or placebo has been administered, subjects will self-report their emotional and cognitive experiences, which are expected to include a subjective feeling of being under the influence of a drug, as well as varying degrees of euphoria and anxiety. We will also evaluate their performance on behavioral tasks, and record physiological data over the course of the sessions. We will then examine the relationship between the response to amphetamine and polymorphisms in the genes selected in the mouse studies. We expect that some of the genes have polymorphisms that are associated with, and may cause, differences in the response to amphetamine among the human volunteers. We believe that differences in the acute response to stimulant drugs like amphetamine, methamphetamine and cocaine are important for determining the genetic risk for the development of drug abuse. Identification of these genes may be useful for developing new therapeutic strategies to treat or prevent drug abuse and addiction.

DESCRIPTION (provided by applicant): This proposal describes a rational system for the identification of genes that control naturally occurring variability for fear conditioning (FC) in mice using a quantitative trait locus (QTL) mapping strategy. As QTL strategies and technologies have matured, it has become increasingly possible to identify the specific genes that cause QTLs. We will take advantage of all available tools, and build on the approaches that we and others have successfully used in previous studies of other phenotypes to identify specific genes that underlie phenotypic variability for FC. In Specific Aim 1 we will examine QTL that we have already discovered in a cross between C57BL/6J (B6) and DBA/2J (D2) mice by employing a panel of B6 x A/J consomic mice. In Specific Aim 2 we will extend the precision of our QTL mapping by using an 8th generation B6 x D2 advanced intercross line (AIL). We will employ both conventional and novel tools for the analysis of the AIL. The novel analysis addresses the problems created by the complex pedigree structure of an AIL by employing techniques originally developed for complex human pedigrees. We will also examine correlations between gene expression and extreme FC phenotype in the final generation of the AIL. In Specific Aim 3 we will examine FC in a panel of inbred mouse strains, which will allow us to employ knowledge about the ancestral single nucleotide polymorphism (SNP) haplotypes that are the basis of genetic diversity in modern laboratory inbred strains. By examining only chromosomal regions already implicated by our preliminary results and the results of the first two Specific Aims, we will mitigate problems that have been associated with this approach in the past. Finally, in Specific Aim 4 we will integrate the information from the prior three Specific Aims and employ a battery of complementary experimental and bioinformatic techniques to identify quantitative trait genes (QTGs), which underlie the QTLs that we have identified. We will utilize knowledge of ancestral SNP haplotypes, clues from the published literature, between-strains coding sequence differences, regional expression data, and strain specific expression polymorphisms in order to identify candidate QTGs. This integrated approach builds on the strategies that have proved successful for gene identification in our own prior studies and the studies of others. Relevance: Anxiety disorders are debilitating illnesses characterized by excessive or inappropriate fears. We will identify genes associated with differential fear learning because evidence form both mice and humans suggests that such differences are genetically associated with anxiety disorders. These genes will later be examined in humans and will advance both diagnosis of anxiety disorders and the development of new and more effective treatments for normal and pathological anxiety.

DESCRIPTION (provided by applicant): Individual differences in the subjective effects of drugs of abuse are partially under genetic control. These differences are believed to influence the risk for drug abuse and dependence. Mice can be used to model sensitivity to the effects of drugs and offer a number of advantages relative to studies in human populations. We are proposing to fine map quantitative trait loci (QTLs) affecting methamphetamine (MA) sensitivity using a 34th generation SM/J x LG/J advanced intercross line (AIL). An AIL is an outbred population that is derived from a cross between two inbred strains and maintained beyond the F2 generation. AILs are powerful populations for fine mapping of QTLs because each individual accumulates a large number of recombinations over the course of many generations. These recombinations break down linkage disequilibrium (LD) between the QTL and surrounding markers, allowing high resolution mapping of the variants that influence quantitative traits. High resolution mapping is critically important to the future of QTL studies. Presently less than 1% of all QTLs identified in mice have led to successful gene identification. This is because standard populations like F2 crosses and small RI panels are good for identifying broad regions that contain QTLs (coarse mapping), but have too much LD to permit fine mapping. Three main factors have precluded the wide-spread use of AIL for fine mapping. First, AIL take several years to create;we have overcome this difficulty by obtaining an AIL that is already at the 34th generation. Second, a large number of markers must be genotyped due to the large number of recombinations;modern high-throughput genotyping makes it possible to type thousands of markers rapidly and at relatively low cost. Third, the methods for the analysis of an AIL have yet to be developed;this proposal addresses this third and final obstacle. AILs have a complex pedigree structure that must be taken into account when analyzing the relationship between genotype and phenotype. No currently available methods properly account for the pedigree structure in these populations, which causes both false negative and false positive errors. We will adapt three variance component methods that are widely used in human genetics. The first method (association) will use a standard mixed model to regress phenotype on QTL genotype while accounting for the correlations due to relatedness among pedigree members. The second method (within-family association) will use the quantitative transmission disequilibrium test (QTDT) to condition on parental genotype, thus eliminating the need to explicitly model the pedigree structure. The third method (linkage) will use classical quantitative trait linkage analysis to model phenotypic similarity using the number of alleles shared identical by descent while explicitly modeling the pedigree structure. We will then use simulations of possible genotypes given the known pedigree structure to evaluate the statistical significance of our results. SM/J and LG/J show extremely divergent locomotor responses to 2 mg/kg MA (>5-fold difference;p<10-13), suggesting the presence of one or more major QTLs for this phenotype.

DESCRIPTION (provided by applicant): Individual differences in the acute response to drugs of abuse can influence future drug taking and may thus influence risk for drug abuse and addiction. Several lines of evidence indicate that genetic variation contributes to variability in the acute response to drugs of abuse. Recently developed genetic methodologies make it possible to identify the underlying genetic factors that govern these differences. We have already collected data from a large number of healthy human subjects, using highly controlled laboratory-based procedures to measure the acute subjective, behavioral, and physiological responses to d-amphetamine. We currently have DNA from 318 such subjects, and will have reached our goal of collecting 400 by the projected start date of this proposal. We have already examined a number of candidate genes using a targeted genotyping strategy that focuses on replicating genetic associations identified by other studies. In this proposal we seek to use a less biased genome-wide approach. Specifically, we will genotype these 400 subjects using the Affymetrix 6.0 array to obtain more than 900,000 single nucleotide polymorphisms (SNPs) as well as information about copy number variants (CNVs). We will approach the analysis of genotypes and phenotypes using a frequentist approach, a modified frequentist approach, and a Bayesian analysis. The first approach will use a standard genome-wide association analysis (GWAS), without any prior hypotheses about which SNPs or CNVs are most likely to be true positives. This approach is risky because our sample is relatively small for such a large number of tests. Nevertheless, the effect size of pharmacogenetic variants may be stronger than for disease traits, so our GWAS may identify genome-wide significant results. Our second approach will consider certain specific subpopulations of SNPs for which we believe there is a strong prior probability of a true positive. We delineate several such subpopulations and discuss their relative strengths and weaknesses. We will incorporate this information by using a weighted multiple testing procedure that facilitates the input of prior information (Roeder et al., 2006, 2007). This method enhances the power to detect associations at candidate loci with minimal loss in power to detect other associations. Our third approach will utilize Bayesian methods, as implemented in BIMBAM (Guan and Stephens 2008). By approaching this same problem with a different statistical framework we can directly estimate the strength of the evidence that a given marker is associated with response to amphetamine. This approach also allows us to account for specific subpopulations that we believe to be enriched for true positives. Our proposal is designed to leverage the substantial investment that has already been made in collecting and carefully phenotyping these 400 subjects. We are taking advantage of technical advances that have made it relatively inexpensive to obtain very large numbers of genotypes. Our statistical approach has the potential to identify SNPs that may contribute to both acute response, and perhaps also other clinically significant endpoints such as drug abuse and addiction. It combines a number of statistical tools to perform a sophisticated GWAS while accounting for other sources of information. We believe that our approach will maximize the potential to obtain new information from our unique sample. PUBLIC HEALTH RELEVANCE: Abuse of stimulant drugs like cocaine, amphetamine and methamphetamine are major public health problems. We are trying to identify genetic differences that influence differences in the response to d- amphetamine, in healthy young adults. About 70% of volunteers like the effects of amphetamine, but about 30% do not;multiple lines of evidence suggest that some of this variability is due to genetic differences. Studies have shown that people who like the effects of a drug continue taking them and are at increased risk for drug abuse. We hope that by identifying genes that influence drug liking we can predict risk of drug abuse and better understand the biological nature of drug abuse in susceptible and resistant people.

DESCRIPTION (provided by applicant): Mice offer a powerful tool for elucidating the genetic basis of behavioral and physiological traits. Reverse genetic approaches such as the creation of knock-out and transgenic mice have been extremely successful in testing hypotheses about the function of specific genes. Forward genetic strategies, which seek to identify the relationship between genes and phenotypes based on standing variation in a heterogeneous population, have also provided insights, but almost never identify the causal genes. This is because traditional approaches to the analysis of quantitative traits in mice are analogous to family-based linkage designs in human, and typically identify large genomic regions that contain many genes. The goal of this proposal is to implement a forward genetic strategy that is similar to human genome-wide association studies (GWAS) and will be able to identify small regions and thus specific genes that are associated with phenotypic variability. We will utilize outbred CD-1 mice because linkage disequilibrium breaks down over short physical distances in these mice compared to other commonly studied mouse populations. Moreover, CD-1 mice are descendants of the same laboratory mice that gave rise to other laboratory strains and are therefore easy to handle and are likely to segregate many of the same alleles. We will exhaustively phenotype 1,008 CD-1 mice for a battery of behavioral and physiological traits. Mice will be genotyped at ~600,000 markers using the new Affymetrix Mouse Diversity Array;we will then perform GWAS to identify QTLs for all phenotypes measured. The traits that we will study are related to psychiatric disease and build on our prior experience in the area of behavioral genetics. The physiological traits are of interest to the diverse array of collaborators that we have assembled. In addition to providing information about these medically important traits, we will demonstrate the broad applicability of this method. We will also employ next-generation sequencing of mRNA (RNASeq) obtained from key brain regions to identify gene expression differences in a subset of these mice. These data will be used to map expression QTLs (eQTLs) and to identify coding polymorphisms. By identifying SNPs that are associated with both behavioral and gene-expression traits we can rapidly identify plausible biological explanations for how these SNPs influence behavior. Such hypotheses are directly testable in mice, which is a major advantage of performing GWAS in mice versus humans. This component of the project will be managed by Dr. Jonathan Pritchard's group, which has conducted similar analyses in human cell lines. In the final phase of this project we will synthesize data about QTLs, eQTLs and coding SNPs. We will implement a Bayesian approach that uses information about eQTLs and coding SNPs as priors for finding QTLs. In addition, we will examine the correlation between gene expression and complex traits in an effort to identify correlations that may not be attributable to a specific genomic locus. The methods proposed in this application are generally applicable to any quantitative trait and have the potential vastly accelerate the process of gene identification, which will aid in the identification of common and rare alleles that contribute to human disease. PUBLIC HEALTH RELEVANCE: Mice are a powerful tool for understanding the genetic basis behavioral and physiological traits;this proposal will use cutting-edge statistical and molecular techniques for this purpose. The results will provide insights into several medically important phenotypes and will also develop new methods for mouse genetics.

DESCRIPTION (provided by applicant): We propose a multidisciplinary, translational genetic study to identify genes that influence sensitivity to the stimulant drugs methamphetamine and amphetamine in mice and humans. We will use mice to better define chromosomal regions (also called quantitative trait loci or QTL) that contain genes that influence sensitivity to the acute locomotor stimulant effects of methamphetamine. This will be accomplished using two complimentary approaches: congenic strains, and advanced intercross lines. Using these techniques, we will be able to determine, with some precision, the location of genes that influence this trait. We will then use a variety of experimental and bioinformatic approaches to parse among the genes located in those regions, which will allow us to identify a small number of high probability candidate genes. We expect that this list of genes will include those responsible for differences in the response to methamphetamine in the mice. We will then explore the effect of polymorphisms in these genes on the subjective response to amphetamine in a sample of carefully screened, healthy human volunteers. All volunteers will give informed consent prior to participation in the study. We will use a double-blind, within-subject design in which each subject receives each dose (placebo, 5,10 and 20 mg) of amphetamine once, in a randomized order, over the course of 4 sessions. During the sessions when drug or placebo has been administered, subjects will self-report their emotional and cognitive experiences, which are expected to include a subjective feeling of being under the influence of a drug, as well as varying degrees of euphoria and anxiety. We will also evaluate their performance on behavioral tasks, and record physiological data over the course of the sessions. We will then examine the relationship between the response to amphetamine and polymorphisms in the genes selected in the mouse studies. We expect that some of the genes have polymorphisms that are associated with, and may cause, differences in the response to amphetamine among the human volunteers. We believe that differences in the acute response to stimulant drugs like amphetamine, methamphetamine and cocaine are important for determining the genetic risk for the development of drug abuse. Identification of these genes may be useful for developing new therapeutic strategies to treat or prevent drug abuse and addiction.

DESCRIPTION (provided by applicant): The Training in Emerging Multidisciplinary Approaches to Mental Health and Disease (TEAM HEAD) represents the evolution of a ten year old training program that has focused on training postdoctoral fellows in the area of psychiatric genetics. This renewal application contains several substantial changes and improvements. First, we are proposing a change of leadership. Specifically, Drs. Abraham Palmer and Nancy Cox (multiple-PIs) will replace Dr. Elliot Gershon, who has served as Program Director for the last 10 years. Dr. Gershon, who is now in his 70s, continues to run an active research lab, but is stepping down as Program Director. He will remain on the executive committee and will provide advice to Drs. Palmer and Cox to assure a smooth transition. Dr. Nancy Cox has been a co-I since the training program's inception and will be particularly involved in determining the program's curriculum; Dr. Palmer will also have significant intellectual input and will be responsible for day-to-day operations. Second, for the past ten years this training program has only had two postdoctoral slots, making it unusually small and preventing it from obtaining a critical mass. Thus, we are proposing to substantially expand the program by increasing the number of postdoctoral slots from 2 to 4. Furthermore, we are adding four slots for graduate students, who will apply for support at the beginning of their second year, once they have joined the lab of one of our training faculty and selected a project that is within the scope of this training program. Third, the focus of the training program has been broadened and refocused with an eye towards meeting the demands that the next generation of scientist will face in the coming decades. For example, there is a much greater emphasis on statistical and computational approaches, which is complimented by close integration with a recently funded Conte Center for Computational Systems Genomics of Neuropsychiatric Phenotypes and the developing Grossman Institute. Additional training faculty, particularly from the Department of Human Genetics, have been recruited to meet the new goals of the program. The overall size of the training faculty has almost doubled as a result of these changes. Students will be supported for 2 years. The number of slots is based on the extremely high quality pool of applicants at both the pre- and post-doctoral levels. Pre-doctoral students will be drawn from a variety of relevant departments and graduate programs. The University of Chicago provides a unique environment with a long history of interdisciplinary collaboration that will benefit the trainees. The scientific interactions among the trainees will be extensive, and will be enhanced by inclusion of a larger cohort of both pre- and post-doctoral students. Formal mechanisms will include student work-in-progress and journal club presentations, trainee- invited speakers, and didactic classes that will be re-structured as full-day retreats and will include a social luncheon. Expanded and appropriate efforts will be made with regard to recruitment of underrepresented minorities and training in ethical conduct of research.

DESCRIPTION (provided by applicant): Among members of the model organism community there is almost universal appreciation that the effect of an allele is frequently modulated by genetic background. This concept has been reinforced by the observation that genetically engineered 'knockout mice' usually have different phenotypes when observed on different genetic backgrounds. Although epistasis is clearly one of the potential explanations for the much discussed 'missing heritability,' human geneticists are mostly focusing their attention on rare alleles instead. This is understandable: an unbiased search for interactions in a typical human genome-wide association study (GWAS) would require about 1012 tests. If the effect sizes of the interactions are even smaller than the main effects that have been detected by GWAS, then power to find such interactions will be very poor. Methods for reducing the number of comparisons by prioritizing certain genes may improve power to detect epistatic interactions. We are proposing to take advantage of the highly structured populations that can be created using mouse crosses to detect epistasis and to map it to specific genetic loci. We are focusing our efforts on the gene Cacna1c, which is among the most robust findings to emerge from genome-wide association studies (GWAS) of psychiatric diseases. CACNA1C has been implicated with genome-wide significance for bipolar disorder, and this finding has been replicated in independent samples. More recent data suggest that CACNA1C may also influence the risk for other psychiatric disorders, including major depression and schizophrenia. Mice that are heterozygous for a null allele of Cacna1c show multiple robust behavioral differences. We are proposing to use Cacna1c mutant mice to test our approach. We will cross an inbred C57BL/6J (B6) mouse that is heterozygous for a knockout of Cacna1c with different inbred strains, producing a cohort of F1s in which half of all offspring will inherit one copy of te null allele (+/-) and the other half will inherit two wild-type alleles (+/+). We will then phenotye these mice for behaviors that are known to be altered in Cacna1c +/- mice. Our goal is to determine which F1 backgrounds are susceptible and which ones are resistant to the null allele for each of several behavioral traits. We will use that data to perform a scan for modifiers. Those modifiers can be tested in the human GWAS datasets that were used for the initial identification of CACNA1C. We will use a mixed model to account for the different degrees of relatedness among the inbred strains by using the program QTLRel, which we have developed for the genetic analysis of complex traits. While the proposed studies focus on Cacna1c, our approach is generally applicable to the detection of epistatic interactions with naturally occurring or engineered mutant alleles, provided that they have a dominant mode of inheritance.

DESCRIPTION (provided by applicant): We will use quantitative genetic techniques to identify genes and alleles that influence a constellation of psychologically complex behavioral phenotypes that are associated with drug abuse. Our center capitalizes on a number of recent advances in study design, next-generation sequencing and statistical methods that together create an exceptional opportunity. Whereas the past two decades have seen enormous advances in both forward (phenotype to genotype) and reverse (genotype to phenotype) genetic studies in mice, there has been much less progress developing and applying the same techniques in rats. Although both forward and reverse genetic studies in mice have been extremely fruitful, many important but complex psychological processes are difficult or impossible to study in mice. For this reason we are proposing to adapt a variety of genetic techniques that we have already used successfully in mice, for behavioral studies in rats. In the current application, we propose to employ state-of-the-art tools to elucidate the genetic basis of a variety of behaviors that are relevant to drug abuse in 4,800 rats. A key strength of our center is that we utilize a unique rat heterogeneous stock (HS). These rats, sometimes referred to as N/NIH, are an HS that was created in 1984 by intercrossing 8 inbred rat strains and have been maintained as an outbred population for 65 generations. This has given rise to numerous accumulated recombinations that make them an ideal resource for performing studies that are analogous to human genome wide association studies (GWAS). Significantly, all 8 inbred founder strains have recently been re-sequenced, identifying 7.2 million SNPs. We will genotype these rats using an innovative next-generation sequencing-based method that provides ~100K SNPs at an extremely low cost. We will use these data to infer haplotypes, which will allow us to impute all 7.2 million SNPs in each rat. In addition to identifying associations between these SNPs and the behavioral traits, we will also examine gene expression in 72 behaviorally naive rats focusing on 4 key brain regions. We will use these data to identify expression quantitative trait loci (eQTLs). We will then integrate all of these data to identify specific genes that influece behavior. Many of the behavioral domains being studied are known to be sexually dimorphic; our study will use both male and female rats, which will allow us to identify sex differences and sex by genotype interactions. We will also identify co-heritability among these traits, as well as pleiotropic effects of individual loci on multiple putatively related behavioral domains. Finally, the proposed center includes numerous educational, career development and public outreach activities. We will implement a program to train high school and undergraduate students. In addition, technicians, graduate students, postdocs and junior faculty will receive career development advice in the form of individual development plans and research performance progress reports. Finally, we will engage in public outreach programs that will draw on diverse communities in Chicago, Buffalo and Memphis.

Project Summary/Abstract (Core C: Sequencing Core) The sequencing core will receive tissue samples from Research Projects 1-3 and use next-generation sequence to produce data that will be analyzed by Research Project 4. Two types of tissue samples will be processed by the core. First, the Sequencing Core will receive spleens from Research Projects 1-3 that will be used for genotyping. Genotype data will be produced using genotype by sequencing (GBS), which is an innovative technology that takes advantage of the rapid advances in next-generation sequencing. This core will extract DNA from the spleens, make GBS libraries, multiplex those libraries at a density of 12 per lane, and sequence them using an Illumina HiSeq 2500. The raw sequencing reads will be sent to Research Project 4 for further analysis. Second, the Sequencing Core will also perform RNA sequencing (RNAseq) of 4 key brain regions: the nucleus accumbens core (AcbC), the lateral habenula (LHb), the infralimic cortex (IL) and the orbitofrontal cortex (OFC). Total RNA from these brain regions will be supplied by Research Project 2, which will use laser capture microdissection (LCM) to accurately dissect these regions from HS rats. This core will perform an amplification step prior to preparing libraries because the amount of mRNA obtained via LCM will be relatively small. We will multiplex those libraries at a density of 5 per lane and sequence them using an Illumina HiSeq 2500. The raw reads will be passed to Research Project 4 for further analysis. The PI's laboratory (Dr. Palmer) has extensive experience with all necessary techniques. This core will have access to numerous next-generation sequencing machines that are already in place at the University of Chicago. In particular, the Department of Human Genetics maintains an Illumina HiSeq 2500; this core will only have to pay for the cost of reagents to use that machine. Dr. Palmer also has access to at least 5 other HiSeq 2500 machines at the University of Chicago, as well as next-generation sequencing machines from other companies (e.g. 454, Pacific Biosystems, etc.). Additional machines are also available at other participating institutions. Because sequencing technology will continue to evolve rapidly, we anticipate that some of the technical details that we describe will change over the proposed funding period; since those changes are difficult to predict we have focused on describing technologies that are currently in use in Dr. Palmer's lab. Note that we are not proposing to purchase any major equipment (e.g. we are not proposing to buy a next- generation sequencing machine) as part of this core. This allows us to deliver the necessary services while minimizing the costs paid by the center. We also discuss the possibility that this core would provide GBS services to outside investigators on a cost recovery basis. Because the analysis of GBS data requires large numbers of samples, such a service would only be appropriate for investigators who are interested in genotyping at least several hundred samples.

Project Summary/Abstract (Core D: Pilot Project Core) The purpose of the Pilot Project Core is to provide resources for follow-up studies that are essential to replicate, validate and extend the results obtained by the center. Follow-up studies may include data collection or analysis. In later years we anticipate conducting follow-up studies in which we produce and phenotype rats with mutations in genes that were identified by our GWAS. Recent advances in the ability to target any gene in the rat genome for deletion (or to produce transgenic overexpressing rats or even knock-ins of putatively causal sequence variants) are one of the key rationales for conducting GWAS in rats. We also hope to use pilot projects to better integrate our findings with emergent findings from human genetic studies. In this section we describe several pilot projects that we would consider for funding and also discuss the mechanism by which we will make decisions about which projects to fund. Pilot project funding will be available to new-, junior-, and senior-level investigators at any of the participating institutions.

Project Summary/Abstract (Research Project 4) In this research project we will analyze behavioral trait, genotype, RNAseq, whole genome sequence data in an effort to identify genes that influence a wide variety of drug abuse related behavioral traits. Genotype data from almost 5,000 heterogeneous stock (HS) rats will be provided by the Sequencing Core (Core C); genotypes will be obtained using an innovative genotype-by-sequencing (GBS) approach. Because the 8 founders of the HS have been fully sequenced, we will be able to impute founder haplotypes and genotypes at 7.2 million SNPs throughout the rat genome. We will use this information to perform a genome-wide association study (GWAS) for each of the behavioral traits measured in Research Projects 1-3. The results of that analysis will identify behavioral quantitative trait loci (QTLs). Because we are also interested in sex differences, we will investigate gene x sex interactions. Furthermore, we will use multitrait mapping methods because many of our behavioral traits are known to be correlated with one another. A key feature of our analysis is that we will account for relatedness among HS individuals. Because they are produced by a large but finite breeding colony, the individuals that we are studying will have varying levels of genetic relationships with one another (e.g. siblings, cousins, etc.). Failure to properly account for such relationships can result in dramatic false positive errors. We have extensive experience using mixed models to deal with similar populations and have developed widely used R packages for this purpose. Another major component of this research project will be the mapping of expression QTL's (eQTLs). As part of Research Project 2, Dr. Hao Chen will collect specific brain regions using laser capture microdissection (LCM). The Sequencing Core (Core C) will perform RNAseq on these brain regions. In this research project we will use RNAseq data to quantify gene expression in each brain region and use these data to map eQTLs. A further goal of this project is to explore genetic correlations between traits, including traits that are not measured in the same rats. Our approach is similar to recent investigations of co-heritability among major psychiatric disorders and depends on the varied levels of relatedness among the HS rats used by Research Projects 1-3. This highly innovative approach will fundamentally enhance our understanding of these drug abuse related behaviors. Finally, we will integrate behavioral QTLs, eQTLs and genome-wide sequence data from the 8 founders of the HS to identify putatively causal genes. Thus, by the end of the proposed five-year project, we expect to identify not only numerous chromosomal regions, but also specific genes within those regions that we believe are causally related to the behavioral QTLs. Subsequent functional studies to replicate and extend these findings using tools such as genetically engineered mutant rats, viral mediated gene transfer and pharmacological manipulations will be funded by the Pilot Project Core (Core D) or by subsequent grants.

PROJECT SUMMARY/ABSTRACT Much of our daily behavior is controlled by stimuli (cues) associated with rewards; these cues can promote either adaptive or maladaptive behavior. We are interested in the role of cues in promoting drug seeking and drug taking behavior because the response to cues may be an essential component of addiction. Cues associated with rewards (conditional stimuli, CSs) powerfully motivate behavior only if they are attributed with incentive motivational properties (incentive salience), and thus acquire the ability to act as incentive stimuli. Among outbred Sprague Dawley rats there is large variation in the propensity of individual rats to attribute incentive salience to food and drug cues. Using a behavioral paradigm called Pavlovian Conditioned Approach (PCA), Sprague Dawley rats can be categorized as sign trackers, meaning that they assign incentive salience to the cue (sign) or goal trackers, which do not assign incentive salience to the cue. This application seeks to take advantage of an outstanding opportunity. As part of an ongoing program project grant, our collaborators are phenotyping several thousand Sprague Dawley rats. They are providing us with phenotypic data and tails as a source of DNA. The current proposal seeks to genotype these samples and then to perform a genome-wide association study. This will allow us to identify genes that underlie the individual variability in PCA observed among outbred Sprague Dawley rats. Because we do not need to pay for the cost of housing or testing the rats, this project is extremely inexpensive yet has the potential to open up completely new molecular avenues for the exploration of the role of cues in shaping behavior. The sample size proposed here (n=3,000) is larger than any similar study that we are aware of in rats or mice and thus provides outstanding power.

Project Summary/Abstract (Core A: Administrative Core) The Administrative Core provides overall coordination, logistical support, and financial accounting for all cores and research projects. Among the tasks the Administrative Core will be responsible for are: coordination and communication with the Internal and External Advisory Boards, bookkeeping, accounting and grants management services; procurement of services, supplies, and equipment; coordination and assistance in preparing progress and fiscal reports; coordination of IACUC protocols and in shipments of rats to research projects 1-3; preservation and dissemination of data generated by the project; organization of events; design and maintenance of a website; providing an initial point of contact for outside researchers wishing to obtain HS rats or GBS services from the relevant cores; conflict resolution; and coordination of all training, mentoring and outreach activities. Education for the next generation of young scientists is a major priority of the proposed Center. Students will be trained in the art and science of interdisciplinary multicenter projects that are required to address our scientific questions. Students at both the high school and undergraduate level will be supported as part of this project. We will utilize existing programs designed to attract and train members of underrepresented minorities, persons from disadvantaged backgrounds and persons with disabilities. In addition, the Administrative Core will also be responsible for the career development of technicians, graduate students, postdocs and junior faculty involved in this project and will use individual development plans and research performance progress reports to achieve these goals. Finally, the Administrative Core will engage in public outreach programs that will draw on the diverse neighborhoods around the University of Chicago.

Project Summary Alcohol use disorders (AUD) place an enormous burden on our society. In addition to their psychological and social toll, it is estimated that AUDs cost the US economy $249 billion in the year 2010 alone. Although there are already several effective behavioral and pharmacological treatments, there is an urgent need for new pharmacotherapies that act via novel mechanisms. We have recently shown that inhibitors of the enzyme Glyoxalase 1 (GLO1) reduce voluntary ethanol drinking in mice (McMurray, et al 2017a). GLO1 is a cytosolic enzyme that metabolizes methylglyoxal (MG). MG is a non-enzymatic side product of glycolysis and is therefore present in all cells. Thus, GLO1 activity is inversely related to MG concentration. We have previously shown that transgenic overexpression of GLO1 increases anxiety-like behavior in mice and decreases MG concentrations in the brain. Reciprocally, we showed that direct administration of MG, genetic knockdown of Glo1 or inhibition of GLO1 using a small molecule inhibitor all decrease anxiety-like behavior and increase MG concentrations in brain. We found that even higher doses of MG produced locomotor depression, ataxia and hypothermia; taken together these data suggested that MG might be acting through GABA-A receptors. Indeed, using a patch clamp procedure, we found that MG is a competitive partial agonist at GABA-A receptors. We also showed that MG is highly selective: it does not activate GABA-B receptors, other ligand gated ion channels or voltage gated ion channels. More recently, we have shown that GLO1 inhibition has antidepressant-like effects, suggesting that GLO1 inhibitors might treat anxiety and depression, both of which are comorbid with AUD. Given the importance of GABA-A signaling in the effects of ethanol, we speculated that GLO1 and MG might also modulate ethanol-related behaviors, which led us to study the effects of GLO1 on ethanol drinking. Those studies showed that inhibition of GLO1 decreased ethanol drinking, which is the rationale for the experiments proposed in this application. We are proposing studies aimed at understanding why inhibition of GLO1 reduces voluntary ethanol drinking. In Aim 1 we will use the intracranial self-stimulation (ICSS) procedure to examine the acute and chronic effects of genetic and pharmacological manipulations of GLO1 on hedonic and anhedonic responses to ethanol in mice. In Aim 2 we will use conditioned place preference (CPP) and conditioned taste aversion (CTA) to study the effects of GLO1 on preference and aversion for ethanol in mice. Finally, in Aim 3, we will use the chronic intermittent ethanol (CIE) procedure to examine the effects of Glo1 inhibitors on acute ethanol withdrawal, compulsive-like responding for ethanol and reinstatement of ethanol seeking behavior after protracted abstinence in rats.

Project Summary (Core C) The purpose of this core is to obtain genotypes for ~5,000 DNA samples derived from rats that are phenotyped by Projects 1, 2, and 3. In addition, we will perform RNA Sequencing (RNASeq) in support of Research Projects 1 and 2. Finally, we use these data to perform genome wide association studies (GWAS) and a number of related analyses. In the prior funding period, Dr. Palmer was the PI of both this core and a research project that developed techniques for genetic analysis. In this renewal application, those two functions have been consolidated into this core. One of the critical technical advances that made this center possible was our development of genotyping- by-sequencing (GBS) for use in rats. Only three SNP genotyping microarrays have ever been developed for use in rats. All three are expensive (>$300 per sample) and none are currently in production. The lack of affordable genotyping platforms was a major impediment to the use of GWAS and related quantitative genetic approaches in rats. In the past funding period, our P50 center has turned the tide by further improving and widely deploying GBS for use in rats. We have continued to refine our GBS techniques, and are now able to obtain ~3.7 million SNPs for less than $50 per sample. Thus, we have increased our output by more than 30-fold while cutting the costs in half. In addition to obtaining genotypes and performing RNASeq, this core will perform genetic analyses, which include GWAS, phenome-wide association analyses (PheWAS), transcriptome wide association analyses (TWAS), heritability estimates, genetic correlations, and a new approach that we call polygenic transcriptomic risk score (PTRS) prediction. All of these analyses are routinely performed in our laboratory and have been the subjects of prior publications that used rats, mice, and humans.

Summary (NOT-OD-21-094 Administrative Supplement to P50DA037844) In this supplement application, we are seeking funds to improve the AI/ML-readiness of data generated by the Center for GWAS in outbred rats. Since the center?s formation in 2014, we have collected extensive data on more than 8,000 genetically unique heterogeneous stock (HS) rats and have secured funding to grow that number to 16,000 by 2025. Data types include genotypes at millions of single nucleotide polymorphisms (SNPs), complex behavioral and physiological phenotypes, RNASeq, ATACSeq, single cell RNASeq, single cell ATACSeq, microbiome, and metabolomic data. While the center is focused on traits relevant to substance abuse, these datasets are much more broadly applicable. They include other behavioral traits relevant to all fields of neuroscience and physiological traits relevant to numerous organ systems and diseases. These data have been carefully curated, including numerous human and automated quality control steps, and are organized as data types available for each unique individual. However, there is no public facing description of the data, and no effort has been put into making them AI/ML-ready. In this proposal, we will improve this situation by bringing together a team with expertise in 1) this specific dataset, 2) best practices for information sharing, and 3) AI/ML for genetic applications. We will begin by bringing the group together to identify the most important and addressable shortcomings. We will then begin to address these goals, meeting frequently to monitor progress and overcome unanticipated challenges. Finally, as the work is completed, our extant network of AI/ML collaborators will perform simple AI/ML exercises to confirm that the improvements are successful. This will be an iterative process; meaning that we may revise specific action items over the course of the project in an effort to maximize impact. We anticipate that improvements will include establishing a website, and making all of our data findable. We will use protocols.io to document each research protocol, will assign RRIDs to all individuals, and will use best practices to make all data FAIR and AI/ML ready. This supplement will provide the impetus and funding to bring together an outstanding team to make sure that NIH?s investment in this unique dataset can be used for cutting edge AI/ML approaches. This project is within the scope of the parent award but does not duplicate any work already supported by the parent grant.

Abstract  Twin studies suggest that approximately 50% of the vulnerability to cocaine use disorder is determined by  genetic factors, but genome-­wide association studies (GWAS) in humans have only begun to identify specific  genes that confer this risk. One major impediment to studies of cocaine use disorder is the complexity of the  phenotype and the lack of control of environmental variables. We propose a complementary approach that  leverages a multidisciplinary, highly collaborative consortium that combines next-­generation sequencing with  state-­of-­the-­art behavioral screening in a unique, genetically diverse, nonhuman animal model. The primary  goal of this proposal is to identify gene variants that are associated with increased vulnerability to compulsive  cocaine use by performing a GWAS in N/NIH heterogeneous stock rats. We will use the most relevant animal  model of cocaine use disorder (i.e., escalation of intravenous cocaine self-­administration) and highly  standardized measures of controlled and compulsive cocaine self-­administration. To increase the impact of  these findings and facilitate translational and basic research studies on the mechanisms underlying compulsive  cocaine use, we will also establish a data/tissue repository (CocaineBioBank.org) from behaviorally and  genetically characterized animals that will allow researchers to further investigate the cellular and molecular  mechanisms underlying compulsive cocaine use and identify the biological changes associated with the  expression of specific gene variants. This project is likely to have a sustained and powerful impact on the field  because it will (1) characterize the transition from controlled to compulsive cocaine use in male and female  outbred rats, (2) identify genes associated with compulsive cocaine use, (3) create the CocainBioBank which  will provide free access to brain, kidney, liver, spleen, ovary, testis, adrenal, and blood samples with a variety  of tissue preservation protocols that will allow the generation of induced pluripotent stem cells as well as  neuroanatomical, molecular, biochemical, and pharmacological studies on behaviorally/genetically  characterized animals.   

PROJECT SUMMARY Genome-wide association studies (GWAS) in model organisms such as mice and rats have identified hundreds of genetic loci that are associated with addictive behaviors, but determining the causal genes remains challenging. Single nucleotide polymorphisms (SNPs) may not adequately tag other classes of variants such as structural and repetitive variants that have higher mutation rates. Thus, both panels of inbred strains and outbred populations will fail to identify a subset of loci and causal alleles. We are proposing to address this serious limitation by using cutting-edge methods to discover and genotype structural variants (SVs) and tandem repeats (TRs). Because they are technically challenging to analyze, SVs and TRs have not yet been adequately surveyed in rodents. However, there is already extensive evidence that SVs and TRs are prevalent in mice and rats, and that they have important functional consequences. Our proposal brings together complementary expertise in human genetics and bioinformatic analysis of SVs (Sebat) and repetitive variation (Gymrek) with established leadership in elucidating the genetic basis of behavioral phenotypes in model organisms (Palmer). This study will generate the first large-scale resource for analyzing the effects of complex variation on mouse and rat phenotypes. We use this resource to examine gene expression and behavioral traits in inbred mice (BXD recombinant inbred strains, the Hybrid Mouse Diversity Panel (HMDP), the Diversity Outbred (DO) mice, the Hybrid Rat Diversity Panel (HRDP) and the Heterogeneous Stock (HS) outbred rats. In Specific Aim 1 we will characterize SVs in inbred and outbred mice and rats by single-molecule sequencing. In Specific Aim 2 we will genotype TR in inbred and outbred mice and rats. Finally, in Specific Aim 3 we will perform GWAS using SV and TR for gene expression and behavioral traits. We will determine the phenotypic consequences of the SV and TR genotypes on gene expression and behavioral traits. We will impute SVs and TRs into outbred mice and rats and then perform GWAS using the wealth of preexisting gene expression and behavioral data that are available for the mouse and rat populations studied in this project. Completion of this project will characterize the SV and TR landscape in mice and rats, elucidate their role in gene expression and complex behavioral traits relevant to addiction, and create a community resource that will enhance numerous ongoing mouse and rat genetic studies.

Project SummaryAbstract Epidemiological studies suggest that genetic variability significantly contributes (20-60%) to the analgesic response to opioids and vulnerability to opioid use disorder. However, genome-wide association studies (GWASs) in humans have only begun to identify specific genes that confer this risk. One major impediment to studies of opioid use disorder is the complexity of the phenotype and lack of control of environmental variables. We propose a complementary approach that leverages a multidisciplinary, highly collaborative consortium that combines next-generation sequencing with state-of-the-art behavioral screening in a unique, genetically diverse, nonhuman animal model. The primary goal of this proposal is to identify gene variants that are associated with greater vulnerability to compulsive oxycodone use, tolerance to the analgesic effects of oxycodone, development of withdrawal-induced hyperalgesia, and sensitivity to FDA-approved medications by performing a GWAS in N/NIH heterogeneous stock rats. We will use the most relevant animal model of oxycodone use disorder (i.e., escalation of intravenous oxycodone self-administration) and highly standardized measures of compulsive oxycodone self-administration combined with longitudinal assessments of pain thresholds. This project is likely to have a sustained and powerful impact on the field because it will (1) characterize the transition from controlled to compulsive oxycodone use and its comorbidity with hyperalgesia in male and female outbred rats, (2) identify genes associated with compulsive oxycodone use, the preclinical efficacy of current medication (e.g., buprenorphine), and the analgesic/hyperalgesic effects of chronic oxycodone use, and (3) facilitate follow-up studies by creating a repository that contains brain and blood with a variety of tissue preservation protocols that will facilitate follow-up and replicative studies by allowing the generation of induced pluripotent stem cells and neuroanatomical, molecular, biochemical, and pharmacological studies on behaviorally/genetically characterized animals.