Building a genetic risk model for bipolar disorder from genome-wide association data with random forest algorithm

Chuang, Li-Chung; Kuo, Po-Hsiu

doi:10.1038/srep39943

Cited by 31 publications

(27 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Total cholesterol, HDL and diabetes local polygenic risk associated with BD overlapped in two neighboring regions on chromosome 3 (3p14.1-p12.3), a region previously found to be associated with hypertension (Koivukoski et al, 2004) that harbors multiple genes known to be involved in HDL and cholesterol traits. This region has also been associated with bipolar disorder, depression and schizophrenia in multiple studies (Chuang and Kuo, 2017;Goes et al, 2015;Rudd et al, 2015; Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). Total cholesterol and LDL local polygenic risk associated with BD overlapped in one region on chromosome 20 (20q11.23-q12), a region previously associated with hypertension and diabetes, although not found to be consistently associated with bipolar disorder.…”

Section: Identifying Pleiotropic Regions Of the Genomementioning

confidence: 99%

Genetic pleiotropy between mood disorders, metabolic, and endocrine traits in a multigenerational pedigree

Kember

Hou²,

Xiao

et al. 2017

Preprint

View full text Add to dashboard Cite

Bipolar disorder (BD) is a mental disorder characterized by alternating periods of depression and mania. Individuals with BD have higher levels of early mortality than the general population, and a substantial proportion of this may be due to increased risk for comorbid diseases. Recent evidence suggests that pleiotropy, either in the form of a single risk-allele or the combination of multiple loci genome-wide, may underlie medical comorbidity between traits and diseases. To identify the molecular events that underlie BD and related medical comorbidities, we generated imputed whole genome sequence (WGS) data using a population specific reference panel, for an extended multigenerational Old Order Amish pedigree (400 family members) segregating BD and related disorders. First, we investigated all putative disease-causing variants at known Mendelian disease loci present in this pedigree. Second, we performed genomic profiling using polygenic risk scores to establish each individual's risk for several complex diseases. To explore the contribution of disease genes to BD we performed gene-based and variant-based association tests for BD, and found that Mendelian disease genes are enriched in the top results from both tests (OR=20.3, p=1×10−3; OR=2.2, p=1×10−2). We next identified a set of Mendelian variants that co-occur in individuals with BD more frequently than their unaffected family members, including the R3527Q mutation inAPOBassociated with hypercholesterolemia. Using polygenic risk scores, we demonstrated that BD individuals from this pedigree were enriched for the same common risk-alleles for BD as in the general population (β=0.416, p=6×10−4). Furthermore, in the extended Amish family we find evidence for a common genetic etiology between BD and clinical autoimmune thyroid disease (p=1×10−4), diabetes (p=1×10−3), and lipid traits such as triglyceride levels (p=3×10−4). We identify genomic regions that contribute to the differences between BD individuals and unaffected family members by calculating local genetic risk for independent LD blocks. Our findings provide evidence for the extensive genetic pleiotropy that can drive epidemiological findings of comorbidities between diseases and other complex traits. Identifying such patterns may enable the subtyping of complex diseases and facilitate our understanding of the genetic mechanisms underlying phenotypic heterogeneity.

show abstract

Section: Identifying Pleiotropic Regions Of the Genomementioning

confidence: 99%

Genetic pleiotropy between mood disorders, metabolic, and endocrine traits in a multigenerational pedigree

Kember

Hou²,

Xiao

et al. 2017

Preprint

View full text Add to dashboard Cite

show abstract

“…Artificial intelligence algorithms, especially machine learning, are being increasingly employed to examine biological data . We started utilizing these for pharmacometric analyses almost 10 years ago .…”

Section: Discussionmentioning

confidence: 99%

Supervised Machine-Learning Reveals That Old and Obese People Achieve Low Dapsone Concentrations

Hall

Pasipanodya

Swancutt

et al. 2017

CPT Pharmacometrics Syst. Pharmacol.

View full text Add to dashboard Cite

The human species is becoming increasingly obese. Dapsone, which is extensively used across the globe for dermatological disorders, arachnid bites, and for treatment of several bacterial, fungal, and parasitic diseases, could be affected by obesity. We performed a clinical experiment, using optimal design, in volunteers weighing 44–150 kg, to identify the effect of obesity on dapsone pharmacokinetic parameters based on maximum‐likelihood solution via the expectation‐maximization algorithm. Artificial intelligence‐based multivariate adaptive regression splines were used for covariate selection, and identified weight and/or age as predictors of absorption, systemic clearance, and volume of distribution. These relationships occurred only between certain patient weight and age ranges, delimited by multiple hinges and regions of discontinuity, not identified by standard pharmacometric approaches. Older and obese people have lower drug concentrations after standard dosing, but with complex patterns. Given that efficacy is concentration‐dependent, optimal dapsone doses need to be personalized for obese patients.

show abstract

“…These results are also compared with the existing related works in order to evaluate the performance of this work, given briefly in Table 4. Machine Learning empowers wide range of applications such as weather forecasting [19] [26], sports success prediction [27] etc. More effective methods improve the predictability of the learning models.…”

Section: Rhinoceros Search Algorithmmentioning

confidence: 99%

“…Two risk genes are identified namely NOG and CTBP1[26]. A genetic risk score model is developed using the RF algorithm to classify between normal and abnormal genes[27].…”

mentioning

confidence: 99%

Diagnostic Gene Biomarker Selection for Alzheimer’s Classification using Machine Learning

Sekaran¹,

Sudha²

2019

IJITEE

View full text Add to dashboard Cite

Alzheimer’s disease (AD), also referred to as Alzheimer’s is a neurodegenerative disease and most common type of dementia. It starts at an older age and slowly progressive over time. It is a brain disease which causes loss of memory, reasoning and thinking capability of a person. Short-term memory loss is one of the main symptoms of the AD. Other common symptoms are said to be mood-swings, difficulty in understanding language and its interpretation etc. The major problem in the AD is, it can’t be reverted, but controllable with proper treatment. Genetic factors have a high impact on developing an AD, which can be inherited through genes. According to recent studies, gene therapy shows better results for Alzheimer’s patients than other common medications. It reduces the risk effect of the AD and has a gradual improvement on the patient’s condition. So, identification of gene biomarkers, having high involvement in developing AD could improve positive response over the treatment. In this paper, gene expressions of AD patients and normal peoples are analyzed using statistical approaches and Machine Learning (ML) algorithms. Differential Gene Expression (DEG) identification has an important part in the selection of most informative genes. Potential gene biomarkers are selected using a meta-heuristic global optimization algorithm called Rhinoceros Search Algorithm (RSA). As an outcome from RSA, 24 novel gene biomarkers are identified. Four supervised ML algorithms such as Support Vector Machines (SVM), Random Forest (RF), Naïve Bayes (NB) and Multilayered Perceptron Neural Network (MLP-NN) are used for the classification of two different group of samples. Among them, RSA-MLP-NN model achieved 100% accuracy on identifying the distinction between AD and normal genes and proved its efficacy.

show abstract

Building a genetic risk model for bipolar disorder from genome-wide association data with random forest algorithm

Cited by 31 publications

References 53 publications

Genetic pleiotropy between mood disorders, metabolic, and endocrine traits in a multigenerational pedigree

Genetic pleiotropy between mood disorders, metabolic, and endocrine traits in a multigenerational pedigree

Supervised Machine-Learning Reveals That Old and Obese People Achieve Low Dapsone Concentrations

Diagnostic Gene Biomarker Selection for Alzheimer’s Classification using Machine Learning

Contact Info

Product

Resources

About