The Translational Machine: A novel machine-learning approach to illuminate complex genetic architectures.

The Translational Machine (TM) is a machine learning (ML)-based analytic pipeline that translates genotypic/variant call data into biologically contextualized features that richly characterize complex variant architectures and permit greater interpretability and biological replication. It also reduces potentially confounding effects of population substructure on outcome prediction. The TM consists of three main components. First, replicable but flexible feature engineering procedures translate genome-scale data into biologically informative features that appropriately contextualize simple variant calls/genotypes within biological and functional contexts. Second, model-free, nonparametric ML-based feature filtering procedures empirically reduce dimensionality and noise of both original genotype calls and engineered features. Third, a powerful ML algorithm for feature selection is used to differentiate risk variant contributions across variant frequency and functional prediction spectra. The TM simultaneously evaluates potential contributions of variants operative under polygenic and heterogeneous models of genetic architecture. Our TM enables integration of biological information (e.g., genomic annotations) within conceptual frameworks akin to geneset-/pathways-based and collapsing methods, but overcomes some of these methods' limitations. The full TM pipeline is executed in R. Our approach and initial findings from its application to a whole-exome schizophrenia case-control data set are presented. These TM procedures extend the findings of the primary investigation and yield novel results.

Genome-wide association study identifies 30 obsessive-compulsive disorder associated loci.

Obsessive-compulsive disorder (OCD) affects ~1% of the population and exhibits a high SNP-heritability, yet previous genome-wide association studies (GWAS) have provided limited information on the genetic etiology and underlying biological mechanisms of the disorder. We conducted a GWAS meta-analysis combining 53,660 OCD cases and 2,044,417 controls from 28 European-ancestry cohorts revealing 30 independent genome-wide significant SNPs and a SNP-based heritability of 6.7%. Separate GWAS for clinical, biobank, comorbid, and self-report sub-groups found no evidence of sample ascertainment impacting our results. Functional and positional QTL gene-based approaches identified 249 significant candidate risk genes for OCD, of which 25 were identified as putatively causal, highlighting WDR6, DALRD3, CTNND1 and genes in the MHC region. Tissue and single-cell enrichment analyses highlighted hippocampal and cortical excitatory neurons, along with D1- and D2-type dopamine receptor-containing medium spiny neurons, as playing a role in OCD risk. OCD displayed significant genetic correlations with 65 out of 112 examined phenotypes. Notably, it showed positive genetic correlations with all included psychiatric phenotypes, in particular anxiety, depression, anorexia nervosa, and Tourette syndrome, and negative correlations with a subset of the included autoimmune disorders, educational attainment, and body mass index.. This study marks a significant step toward unraveling its genetic landscape and advances understanding of OCD genetics, providing a foundation for future interventions to address this debilitating disorder.

Trans-diagnostic determinants of psychotherapeutic treatment response: The pressing need and new opportunities for a more systematic way of selecting psychotherapeutic treatment in the age of virtual service delivery.

Numerous forms of psychotherapy have demonstrated effectiveness for individuals with specific mental disorders. It is, therefore, the task of the clinician to choose the most appropriate therapeutic approach for any given client to maximize effectiveness. This can prove to be a difficult task due to at least three considerations: (1) there is no treatment approach, method or model that works well on all patients, even within a particular diagnostic class; (2) several treatments are equally efficacious (i.e., more likely to be effective than no treatment at all) when considered only in terms of the patient's diagnosis; and (3) effectiveness in the real-world therapeutic setting is determined by a host of non-diagnostic factors. Typically, consideration of these latter, trans-diagnostic factors is unmethodical or altogether excluded from treatment planning - often resulting in suboptimal patient care, inappropriate clinic resource utilization, patient dissatisfaction with care, patient demoralization/hopelessness, and treatment failure. In this perspective article, we argue that a more systematic research on and clinical consideration of trans-diagnostic factors determining psychotherapeutic treatment outcome (i.e., treatment moderators) would be beneficial and - with the seismic shift toward online service delivery - is more feasible than it used to be. Such a transition toward more client-centered care - systematically considering variables such as sociodemographic characteristics, patient motivation for change, self-efficacy, illness acuity, character pathology, trauma history when making treatment choices - would result in not only decreased symptom burden and improved quality of life but also better resource utilization in mental health care and improved staff morale reducing staff burnout and turnover.

The evolution of deep brain stimulation for neuropsychiatric disorders.

Deep brain stimulation (DBS) is the most focal method for stimulating the human brain. In contrast to lesions, DBS is nonablative, with the advantages of reversibility and adjustability. Thus, therapeutic effectiveness can be enhanced and stimulation-related side effects minimized during long-term patient management. While DBS is an approved adjunct therapy for severe, medication-refractory movement disorders, it remains investigational in neuropsychiatry. However, experience to date, though limited, suggests that DBS may offer a degree of hope to patients with severe and treatment-resistant neuropsychiatric illness. Thus far, work in obsessive-compulsive disorder (OCD), the first psychiatric condition studied using modern DBS devices, has shown consistently positive results across multiple small-scale studies. Work in treatment-resistant Major Depressive Disorder (MDD) also suggests therapeutic potential in preliminary studies, generating cautious optimism for this indication. With the increase in potential applications, a number of clinical and preclinical research efforts have now focused on understanding the mechanisms of action of DBS. Further development of DBS for these and other illnesses with primarily behavioral symptoms will require thoughtful collaboration among multiple disciplines.

Determinants of prostate cancer stage in northern New England: USA Franco-American contextual effects.

Despite screening for prostate cancer, mortality in the United States remains substantial. In northern New England, we know little about either determinants of stage at diagnosis--an important predictor of survival--or health outcomes for Franco-Americans, the region's largest ethnic minority. The objective of this investigation was to identify predictors of late prostate cancer stage in a rural, predominantly white state with a large Franco-American population. The Maine Cancer Registry provided incident cases from 1995 to 1998. We modeled individual-level variables (age, sex, race, French ethnicity by surname, and payer) and contextual/town-level variables (socioeconomic measures, population density, Franco ancestry proportion, distance to health care, and weather severity) with multiple logistic regression for late stage. We found that age categories 50-64, 65-74, and 75-84 years--but not 40-49 years--(versus 85+) were protective for late stage, as was residence in higher snowfall areas. Diagnosis in the earlier years of the study, particularly for French-surnamed men, and residence in a high-Franco area conferred greater risk for late disease. However, in a two-way interaction, residence in towns with high Franco ancestry proportion protected French-surnamed men (OR=0.09, type 3 p<0.0593). Using an established framework for social network theory we explore the potential reasons for this interaction, including: high social cohesion, a wide range of strong ties of long duration, and frequent contact, which might have facilitated access to resources as well as social support and normative influences toward health care seeking. The absence of an association of cancer stage with socioeconomic variables may stem from the mixed sociodemographic profiles in rural and urban regions of Maine. We feel that further research should therefore refine these and other contextual measures to elucidate effects on preventable morbidity and mortality; expand our knowledge of Franco-American health outcomes and social networks; and evaluate the utility of assigning French ethnicity by surname.

Cancer of the colorectum in Maine, 1995-1998: determinants of stage at diagnosis in a rural state.

CONTEXT: Despite screening for colorectal cancer, mortality in the United States remains substantial. In northern New England, little is known about predictors of stage at diagnosis, an important determinant of survival and mortality. PURPOSE: The objective of this study was to identify predictors of late stage at diagnosis for colorectal cancer in a rural state with a predominantly white population and a large Franco-American minority. METHODS: Incident cases from 1995-1998 were obtained from the Maine Cancer Registry. Individual-level variables (age, sex, race, French ethnicity by surname, and payer) and contextual/town-level variables (socioeconomic status, population density, Franco ancestry proportion, distance to health care, and weather) were modeled with multiple logistic regression for late stage. FINDINGS: Increasing distance to primary care provider was associated with late stage for colorectal cancer. Compared to patients aged > or =85 years, those aged 65-84 years were less likely to be diagnosed late, while those aged 35-49 years were more likely--although not significantly--to have late stage at diagnosis. Associations were not found with socioeconomic variables. CONCLUSIONS: The finding regarding distance to primary care may be consistent with studies showing that rurality and distance to care predict reduced utilization of health care services and worse health outcomes. The finding regarding age has implications for the education of younger high-risk patients and their physicians. The absence of positive findings with regard to socioeconomic variables may stem from the uniquely mixed sociodemographic profiles in rural and urban regions of Maine. Further research should refine these and other contextual measures to elucidate effects on rural health and should further evaluate the utility of assigning French ethnicity by surname in order to identify health disparities.

Genome Wide Association Study (GWAS) between Attention Deficit Hyperactivity Disorder (ADHD) and Obsessive Compulsive Disorder (OCD).

Objective: The aim of this study was to identify any potential genetic overlap between attention deficit hyperactivity disorder (ADHD) and obsessive compulsive disorder (OCD). We hypothesized that since these disorders share a sub-phenotype, they may share common risk alleles. In this manuscript, we report the overlap found between these two disorders. Methods: A meta-analysis was conducted between ADHD and OCD, and polygenic risk scores (PRS) were calculated for both disorders. In addition, a protein-protein analysis was completed in order to examine the interactions between proteins; p-values for the protein-protein interaction analysis was calculated using permutation. Conclusion: None of the single nucleotide polymorphisms (SNPs) reached genome wide significance and there was little evidence of genetic overlap between ADHD and OCD.

Prediction of remission in obsessive compulsive disorder using a novel machine learning strategy.

The study objective was to apply machine learning methodologies to identify predictors of remission in a longitudinal sample of 296 adults with a primary diagnosis of obsessive compulsive disorder (OCD). Random Forests is an ensemble machine learning algorithm that has been successfully applied to large-scale data analysis across vast biomedical disciplines, though rarely in psychiatric research or for application to longitudinal data. When provided with 795 raw and composite scores primarily from baseline measures, Random Forest regression prediction explained 50.8% (5000-run average, 95% bootstrap confidence interval [CI]: 50.3-51.3%) of the variance in proportion of time spent remitted. Machine performance improved when only the most predictive 24 items were used in a reduced analysis. Consistently high-ranked predictors of longitudinal remission included Yale-Brown Obsessive Compulsive Scale (Y-BOCS) items, NEO items and subscale scores, Y-BOCS symptom checklist cleaning/washing compulsion score, and several self-report items from social adjustment scales. Random Forest classification was able to distinguish participants according to binary remission outcomes with an error rate of 24.6% (95% bootstrap CI: 22.9-26.2%). Our results suggest that clinically-useful prediction of remission may not require an extensive battery of measures. Rather, a small set of assessment items may efficiently distinguish high- and lower-risk patients and inform clinical decision-making.

Gene ontology analysis of pairwise genetic associations in two genome-wide studies of sporadic ALS.

BACKGROUND: It is increasingly clear that common human diseases have a complex genetic architecture characterized by both additive and nonadditive genetic effects. The goal of the present study was to determine whether patterns of both additive and nonadditive genetic associations aggregate in specific functional groups as defined by the Gene Ontology (GO). RESULTS: We first estimated all pairwise additive and nonadditive genetic effects using the multifactor dimensionality reduction (MDR) method that makes few assumptions about the underlying genetic model. Statistical significance was evaluated using permutation testing in two genome-wide association studies of ALS. The detection data consisted of 276 subjects with ALS and 271 healthy controls while the replication data consisted of 221 subjects with ALS and 211 healthy controls. Both studies included genotypes from approximately 550,000 single-nucleotide polymorphisms (SNPs). Each SNP was mapped to a gene if it was within 500 kb of the start or end. Each SNP was assigned a p-value based on its strongest joint effect with the other SNPs. We then used the Exploratory Visual Analysis (EVA) method and software to assign a p-value to each gene based on the overabundance of significant SNPs at the α = 0.05 level in the gene. We also used EVA to assign p-values to each GO group based on the overabundance of significant genes at the α = 0.05 level. A GO category was determined to replicate if that category was significant at the α = 0.05 level in both studies. We found two GO categories that replicated in both studies. The first, 'Regulation of Cellular Component Organization and Biogenesis', a GO Biological Process, had p-values of 0.010 and 0.014 in the detection and replication studies, respectively. The second, 'Actin Cytoskeleton', a GO Cellular Component, had p-values of 0.040 and 0.046 in the detection and replication studies, respectively. CONCLUSIONS: Pathway analysis of pairwise genetic associations in two GWAS of sporadic ALS revealed a set of genes involved in cellular component organization and actin cytoskeleton, more specifically, that were not reported by prior GWAS. However, prior biological studies have implicated actin cytoskeleton in ALS and other motor neuron diseases. This study supports the idea that pathway-level analysis of GWAS data may discover important associations not revealed using conventional one-SNP-at-a-time approaches.

Comparative linkage meta-analysis reveals regionally-distinct, disparate genetic architectures: application to bipolar disorder and schizophrenia.

New high-throughput, population-based methods and next-generation sequencing capabilities hold great promise in the quest for common and rare variant discovery and in the search for "missing heritability." However, the optimal analytic strategies for approaching such data are still actively debated, representing the latest rate-limiting step in genetic progress. Since it is likely a majority of common variants of modest effect have been identified through the application of tagSNP-based microarray platforms (i.e., GWAS), alternative approaches robust to detection of low-frequency (1-5% MAF) and rare (<1%) variants are of great importance. Of direct relevance, we have available an accumulated wealth of linkage data collected through traditional genetic methods over several decades, the full value of which has not been exhausted. To that end, we compare results from two different linkage meta-analysis methods--GSMA and MSP--applied to the same set of 13 bipolar disorder and 16 schizophrenia GWLS datasets. Interestingly, we find that the two methods implicate distinct, largely non-overlapping, genomic regions. Furthermore, based on the statistical methods themselves and our contextualization of these results within the larger genetic literatures, our findings suggest, for each disorder, distinct genetic architectures may reside within disparate genomic regions. Thus, comparative linkage meta-analysis (CLMA) may be used to optimize low-frequency and rare variant discovery in the modern genomic era.