Systems genetics studies for unraveling genetic basis of complex traits have been one of the most propitious research area with the advance of high-throughput biotechnologies. This thesis presents several computational and statistical challenges in effective design and analysis of systems genetics studies and present novel methodological advances and corresponding results in several specific contexts of systems genetics studies. First, I present an extensive haplotype analysis on a recently collected catalogue of genetic variation among inbred mouse strains, which revealed the contribution from ancestral subspecies, haplotype block structure, and complex history of each genomic segments among the inbred mouse strains. In addition, I accurately imputed the uncollected genotypes in the resource by developing a novel and efficient genotype imputation method which adaptively learns parameters from data using an Expectaion-Maximuzation (EM) algorithm. Our method is demonstrated to outperform previous methods in both mouse and human data. Statistical analyses in systems genetics studies are often confounded by unmodeled factors such as heterogeneous sample structure. Recent studies suggested that mixed models correct for the sample structure in association mapping, but the available methods suffer from substantial computational cost to be applied in genome- wide association mapping. I developed the Efficient Mixed Model Association (EMMA), which takes advantage of the invariant structure of eigenvectors in applying mixed models for association mapping, which substantially increase the computational efficiency in several orders of magnitude. Our method was shown to successfully reduce inflated false positives in in silico genome-wide association mapping of inbred mouse strains involving hundreds of thousands of markers. I further extend EMMA to accommodate even larger scale of genome-wide association mapping in humans, typically involving several thousands or more individuals, and demonstrate that the method consistently eliminates the significant over-dispersion of test statistics across multiple human data sets. The method has been further employed in correcting for a different type of confounding effects in expression studies. I developed a novel mixed-model method that corrects for the spurious associations and trans- regulatory bands caused by systematic confounding effects using inter-sample correlation of expression measurements. Finally, in the design of association studies using inbred strains, I propose a novel trait mapping strategy using hybrid mouse diversity panel (HMDP). By integrating classical inbreds and multiple sets of recombinant inbreds while precisely accounting for the sample structure using high-density markers with EMMA, the proposed design is shown to much more powerfully and precisely identify previously known associations than previous approaches.

Admixture-the mixing of genomes from divergent populations-is increasingly appreciated as a central process in evolution. To characterize and quantify patterns of admixture across the genome, a number of methods have been developed for local ancestry inference. However, existing approaches have a number of shortcomings. First, all local ancestry inference methods require some prior assumption about the expected ancestry tract lengths. Second, existing methods generally require genotypes, which is not feasible to obtain for many next-generation sequencing projects. Third, many methods assume samples are diploid, however a wide variety of sequencing applications will fail to meet this assumption. To address these issues, we introduce a novel hidden Markov model for estimating local ancestry that models the read pileup data, rather than genotypes, is generalized to arbitrary ploidy, and can estimate the time since admixture during local ancestry inference. We demonstrate that our method can simultaneously estimate the time since admixture and local ancestry with good accuracy, and that it performs well on samples of high ploidy-i.e. 100 or more chromosomes. As this method is very general, we expect it will be useful for local ancestry inference in a wider variety of populations than what previously has been possible. We then applied our method to pooled sequencing data derived from populations of Drosophila melanogaster on an ancestry cline on the east coast of North America. We find that regions of local recombination rates are negatively correlated with the proportion of African ancestry, suggesting that selection against foreign ancestry is the least efficient in low recombination regions. Finally we show that clinal outlier loci are enriched for genes associated with gene regulatory functions, consistent with a role of regulatory evolution in ecological adaptation of admixed D. melanogaster populations. Our results illustrate the potential of local ancestry inference for elucidating fundamental evolutionary processes.

With the development of high-throughput sequencing and genotyping technologies, the number of markers collected in genetic association studies is growing rapidly, increasing the importance of methods for correcting for multiple hypothesis testing. The permutation test is widely considered the gold standard for accurate multiple testing correction, but it is often computationally impractical for these large datasets. Recently, several studies proposed efficient alternative approaches to the permutation test based on the multivariate normal distribution (MVN). However, they cannot accurately correct for multiple testing in genome-wide association studies for two reasons. First, these methods require partitioning of the genome into many disjoint blocks and ignore all correlations between markers from different blocks. Second, the true null distribution of the test statistic often fails to follow the asymptotic distribution at the tails of the distribution. We propose an accurate and efficient method for multiple testing correction in genome-wide association studies--SLIDE. Our method accounts for all correlation within a sliding window and corrects for the departure of the true null distribution of the statistic from the asymptotic distribution. In simulations using the Wellcome Trust Case Control Consortium data, the error rate of SLIDE's corrected p-values is more than 20 times smaller than the error rate of the previous MVN-based methods' corrected p-values, while SLIDE is orders of magnitude faster than the permutation test and other competing methods. We also extend the MVN framework to the problem of estimating the statistical power of an association study with correlated markers and propose an efficient and accurate power estimation method SLIP. SLIP and SLIDE are available at http://slide.cs.ucla.edu.

β-blockers and β-agonists are primarily used to treat cardiovascular diseases. Inter-individual variability in response to both drug classes is well recognized, yet the identity and relative contribution of the genetic players involved are poorly understood. This work is the first genome-wide association study (GWAS) addressing the values and susceptibility of cardiovascular-related traits to a selective β(1)-blocker, Atenolol (ate), and a β-agonist, Isoproterenol (iso). The phenotypic dataset consisted of 27 highly heritable traits, each measured across 22 inbred mouse strains and four pharmacological conditions. The genotypic panel comprised 79922 informative SNPs of the mouse HapMap resource. Associations were mapped by Efficient Mixed Model Association (EMMA), a method that corrects for the population structure and genetic relatedness of the various strains. A total of 205 separate genome-wide scans were analyzed. The most significant hits include three candidate loci related to cardiac and body weight, three loci for electrocardiographic (ECG) values, two loci for the susceptibility of atrial weight index to iso, four loci for the susceptibility of systolic blood pressure (SBP) to perturbations of the β-adrenergic system, and one locus for the responsiveness of QTc (p<10(-8)). An additional 60 loci were suggestive for one or the other of the 27 traits, while 46 others were suggestive for one or the other drug effects (p<10(-6)). Most hits tagged unexpected regions, yet at least two loci for the susceptibility of SBP to β-adrenergic drugs pointed at members of the hypothalamic-pituitary-thyroid axis. Loci for cardiac-related traits were preferentially enriched in genes expressed in the heart, while 23% of the testable loci were replicated with datasets of the Mouse Phenome Database (MPD). Altogether these data and validation tests indicate that the mapped loci are relevant to the traits and responses studied.

Droplet single-cell RNA-sequencing (dscRNA-seq) has enabled rapid, massively parallel profiling of transcriptomes. However, assessing differential expression across multiple individuals has been hampered by inefficient sample processing and technical batch effects. Here we describe a computational tool, demuxlet, that harnesses natural genetic variation to determine the sample identity of each droplet containing a single cell (singlet) and detect droplets containing two cells (doublets). These capabilities enable multiplexed dscRNA-seq experiments in which cells from unrelated individuals are pooled and captured at higher throughput than in standard workflows. Using simulated data, we show that 50 single-nucleotide polymorphisms (SNPs) per cell are sufficient to assign 97% of singlets and identify 92% of doublets in pools of up to 64 individuals. Given genotyping data for each of eight pooled samples, demuxlet correctly recovers the sample identity of >99% of singlets and identifies doublets at rates consistent with previous estimates. We apply demuxlet to assess cell-type-specific changes in gene expression in 8 pooled lupus patient samples treated with interferon (IFN)-β and perform eQTL analysis on 23 pooled samples.

Significant advances have been made in the discovery of genes affecting bone mineral density (BMD); however, our understanding of its genetic basis remains incomplete. In the current study, genome-wide association (GWA) and co-expression network analysis were used in the recently described Hybrid Mouse Diversity Panel (HMDP) to identify and functionally characterize novel BMD genes. In the HMDP, a GWA of total body, spinal, and femoral BMD revealed four significant associations (-log10P>5.39) affecting at least one BMD trait on chromosomes (Chrs.) 7, 11, 12, and 17. The associations implicated a total of 163 genes with each association harboring between 14 and 112 genes. This list was reduced to 26 functional candidates by identifying those genes that were regulated by local eQTL in bone or harbored potentially functional non-synonymous (NS) SNPs. This analysis revealed that the most significant BMD SNP on Chr. 12 was a NS SNP in the additional sex combs like-2 (Asxl2) gene that was predicted to be functional. The involvement of Asxl2 in the regulation of bone mass was confirmed by the observation that Asxl2 knockout mice had reduced BMD. To begin to unravel the mechanism through which Asxl2 influenced BMD, a gene co-expression network was created using cortical bone gene expression microarray data from the HMDP strains. Asxl2 was identified as a member of a co-expression module enriched for genes involved in the differentiation of myeloid cells. In bone, osteoclasts are bone-resorbing cells of myeloid origin, suggesting that Asxl2 may play a role in osteoclast differentiation. In agreement, the knockdown of Asxl2 in bone marrow macrophages impaired their ability to form osteoclasts. This study identifies a new regulator of BMD and osteoclastogenesis and highlights the power of GWA and systems genetics in the mouse for dissecting complex genetic traits.

Background

In cystic fibrosis (CF), the spectrum and frequency of CFTR variants differ by geography and race/ethnicity. CFTR variants in White patients are well-described compared with Latino patients. No studies of CFTR variants have been done in patients with CF in the Dominican Republic or Puerto Rico.

Methods

CFTR was sequenced in 61 Dominican Republican patients and 21 Puerto Rican patients with CF and greater than ​​​​60 mmol/L sweat chloride. The spectrum of CFTR variants was identified and the proportion of patients with 0, 1, or 2 CFTR variants identified was determined. The functional effects of identified CFTR variants were investigated using clinical annotation databases and computational prediction tools.

Results

Our study found 10% of Dominican patients had two CFTR variants identified compared with 81% of Puerto Rican patients. No CFTR variants were identified in 69% of Dominican patients and 10% of Puerto Rican patients. In Dominican patients, there were 19 identified CFTR variants, accounting for 25 out of 122 disease alleles (20%). In Puerto Rican patients, there were 16 identified CFTR variants, accounting for 36 out of 42 disease alleles (86%) in Puerto Rican patients. Thirty CFTR variants were identified overall. The most frequent variants for Dominican patients were p.Phe508del and p.Ala559Thr and for Puerto Rican patients were p.Phe508del, p.Arg1066Cys, p.Arg334Trp, and p.I507del.

Conclusions

In this first description of the CFTR variants in patients with CF from the Dominican Republic and Puerto Rico, there was a low detection rate of two CFTR variants after full sequencing with the majority of patients from the Dominican Republic without identified variants.

To identify genetic determinants of airway dysfunction, we performed a transcriptome-wide association study for asthma by combining RNA-seq data from the nasal airway epithelium of 681 children, with UK Biobank genetic association data. Our airway analysis identified 95 asthma genes, 58 of which were not identified by transcriptome-wide association analyses using other asthma-relevant tissues. Among these genes were MUC5AC, an airway mucin, and FOXA3, a transcriptional driver of mucus metaplasia. Muco-ciliary epithelial cultures from genotyped donors revealed that the MUC5AC risk variant increases MUC5AC protein secretion and mucus secretory cell frequency. Airway transcriptome-wide association analyses for mucus production and chronic cough also identified MUC5AC. These cis-expression variants were associated with trans effects on expression; the MUC5AC variant was associated with upregulation of non-inflammatory mucus secretory network genes, while the FOXA3 variant was associated with upregulation of type-2 inflammation-induced mucus-metaplasia pathway genes. Our results reveal genetic mechanisms of airway mucus pathobiology.

Background

Mitochondria support critical cellular functions, such as energy production through oxidative phosphorylation, regulation of reactive oxygen species, apoptosis, and calcium homeostasis.

Objective

Given the heightened level of cellular activity in patients with asthma, we sought to determine whether mitochondrial DNA (mtDNA) copy number measured in peripheral blood differed between individuals with and without asthma.

Methods

Whole genome sequence data was generated as part of the Trans-Omics for Precision Medicine (TOPMed) Program on participants from the Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity (SAPPHIRE) and the Study of African Americans, Asthma, Genes, & Environment II (SAGE II). We restricted our analysis to individuals who self-identified as African American (3,651 asthma cases and 1,344 controls). Mitochondrial copy number was estimated using the sequencing read depth ratio for the mitochondrial and nuclear genomes. Respiratory complex expression was assessed using RNA-sequencing.

Results

Average mitochondrial copy number was significantly higher among individuals with asthma when compared with controls (SAPPHIRE: 218.60 vs. 200.47, P<0.001; SAGE II: 235.99 vs. 223.07, P<0.001). Asthma status was significantly associated with mitochondrial copy number after accounting for potential explanatory variables, such as participant age, sex, leukocyte counts, and mitochondrial haplogroup. Despite the consistent relationship between asthma status and mitochondrial copy number, the latter was not associated with time-to-exacerbation or patient-reported asthma control. Mitochondrial respiratory complex gene expression was disproportionately lower in individuals with asthma when compared with individuals without asthma and other protein-encoding genes.

Conclusions

We observed a robust association between asthma and higher mitochondrial copy number. Asthma having an effect on mitochondria function was also supported by lower respiratory complex gene expression in this group.

Rationale: The 17q12-21.1 locus is one of the most highly replicated genetic associations with asthma. Individuals of African descent have lower linkage disequilibrium in this region, which could facilitate identifying causal variants.Objectives: To identify functional variants at 17q12-21.1 associated with early-onset asthma among African American individuals.Methods: We evaluated African American participants from SAPPHIRE (Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-Ethnicity) (n = 1,940), SAGE II (Study of African Americans, Asthma, Genes and Environment) (n = 885), and GCPD-A (Study of the Genetic Causes of Complex Pediatric Disorders-Asthma) (n = 2,805). Associations with asthma onset at ages under 5 years were meta-analyzed across cohorts. The lead signal was reevaluated considering haplotypes informed by genetic ancestry (i.e., African vs. European). Both an expression-quantitative trait locus analysis and a phenome-wide association study were performed on the lead variant.Measurements and Main Results: The meta-analyzed results from SAPPHIRE, SAGE II, and the GCPD-A identified rs11078928 as the top association for early-onset asthma. A haplotype analysis suggested that the asthma association partitioned most closely with the rs11078928 genotype. Genetic ancestry did not appear to influence the effect of this variant. In the expression-quantitative trait locus analysis, rs11078928 was related to alternative splicing of GSDMB (gasdermin-B) transcripts. The phenome-wide association study of rs11078928 suggested that this variant was predominantly associated with asthma and asthma-associated symptoms.Conclusions: A splice-acceptor polymorphism appears to be a causal variant for asthma at the 17q12-21.1 locus. This variant appears to have the same magnitude of effect in individuals of African and European descent.