Dataset Information

Meta-analytic framework for sparse K-means to identify disease subtypes in multiple transcriptomic studies.

ABSTRACT: Disease phenotyping by omics data has become a popular approach that potentially can lead to better personalized treatment. Identifying disease subtypes via unsupervised machine learning is the first step towards this goal. In this paper, we extend a sparse K-means method towards a meta-analytic framework to identify novel disease subtypes when expression profiles of multiple cohorts are available. The lasso regularization and meta-analysis identify a unique set of gene features for subtype characterization. An additional pattern matching reward function guarantees consistent subtype signatures across studies. The method was evaluated by simulations and leukemia and breast cancer data sets. The identified disease subtypes from meta-analysis were characterized with improved accuracy and stability compared to single study analysis. The breast cancer model was applied to an independent METABRIC dataset and generated improved survival difference between subtypes. These results provide a basis for diagnosis and development of targeted treatments for disease subgroups.

SUBMITTER: Huo Z

PROVIDER: S-EPMC4908837 | biostudies-literature | 2016

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Meta-analytic framework for sparse <i>K</i>-means to identify disease subtypes in multiple transcriptomic studies.

Huo Zhiguang Z Ding Ying Y Liu Silvia S Oesterreich Steffi S Tseng George G

Journal of the American Statistical Association 20160505 513

Disease phenotyping by omics data has become a popular approach that potentially can lead to better personalized treatment. Identifying disease subtypes via unsupervised machine learning is the first step towards this goal. In this paper, we extend a sparse <i>K</i>-means method towards a meta-analytic framework to identify novel disease subtypes when expression profiles of multiple cohorts are available. The lasso regularization and meta-analysis identify a unique set of gene features for subty ...[more]

PMID: 27330233

Similar Datasets

Project description:BACKGROUND:The increasing use of engineered nanomaterials (ENMs) of varying physical and chemical characteristics poses a great challenge for screening and assessing the potential pathology induced by these materials, necessitating novel toxicological approaches. Toxicogenomics measures changes in mRNA levels in cells and tissues following exposure to toxic substances. The resulting information on altered gene expression profiles, associated pathways, and the doses at which these changes occur, are used to identify the underlying mechanisms of toxicity and to predict disease outcomes. We evaluated the applicability of toxicogenomics data in identifying potential lung-specific (genomic datasets are currently available from experiments where mice have been exposed to various ENMs through this common route of exposure) disease outcomes following exposure to ENMs. METHODS:Seven toxicogenomics studies describing mouse pulmonary responses over time following intra-tracheal exposure to increasing doses of carbon nanotubes (CNTs), carbon black, and titanium dioxide (TiO2) nanoparticles of varying properties were examined to understand underlying mechanisms of toxicity. mRNA profiles from these studies were compared to the publicly available datasets of 15 other mouse models of lung injury/diseases induced by various agents including bleomycin, ovalbumin, TNF?, lipopolysaccharide, bacterial infection, and welding fumes to delineate the implications of ENM-perturbed biological processes to disease pathogenesis in lungs. RESULTS:The meta-analysis revealed two distinct clusters-one driven by TiO2 and the other by CNTs. Unsupervised clustering of the genes showing significant expression changes revealed that CNT response clustered with bleomycin injury and bacterial infection models, both of which are known to induce lung fibrosis, in a post-exposure-time dependent manner, irrespective of the CNT's physical-chemical properties. TiO2 samples clustered separately from CNTs and disease models. CONCLUSIONS:These results indicate that in the absence of apical toxicity data, a tiered strategy beginning with short term, in vivo tissue transcriptomics profiling can effectively and efficiently screen new ENMs that have a higher probability of inducing pulmonary pathogenesis.

Project description:The identification of phenotypic associations in high-dimensional brain connectivity data represents the next frontier in the neuroimaging connectomics era. Exploration of brain-phenotype relationships remains limited by statistical approaches that are computationally intensive, depend on a priori hypotheses, or require stringent correction for multiple comparisons. Here, we propose a computationally efficient, data-driven technique for connectome-wide association studies (CWAS) that provides a comprehensive voxel-wise survey of brain-behavior relationships across the connectome; the approach identifies voxels whose whole-brain connectivity patterns vary significantly with a phenotypic variable. Using resting state fMRI data, we demonstrate the utility of our analytic framework by identifying significant connectivity-phenotype relationships for full-scale IQ and assessing their overlap with existent neuroimaging findings, as synthesized by openly available automated meta-analysis (www.neurosynth.org). The results appeared to be robust to the removal of nuisance covariates (i.e., mean connectivity, global signal, and motion) and varying brain resolution (i.e., voxelwise results are highly similar to results using 800 parcellations). We show that CWAS findings can be used to guide subsequent seed-based correlation analyses. Finally, we demonstrate the applicability of the approach by examining CWAS for three additional datasets, each encompassing a distinct phenotypic variable: neurotypical development, Attention-Deficit/Hyperactivity Disorder diagnostic status, and L-DOPA pharmacological manipulation. For each phenotype, our approach to CWAS identified distinct connectome-wide association profiles, not previously attainable in a single study utilizing traditional univariate approaches. As a computationally efficient, extensible, and scalable method, our CWAS framework can accelerate the discovery of brain-behavior relationships in the connectome.

Project description:SignificanceConcerns about the reproducibility of experimental findings have recently emerged in many disciplines, from psychology to medicine and neuroscience. As NIRS is a relatively recent brain imaging technique, the question of reproducibility has not yet been systematically addressed.AimThe current study seeks to test the replicability of effects observed in NIRS experiments assessing young infants' rule-learning ability.ApproachWe conducted meta-analyses and mixed-effects modeling-based inferential statistics to determine whether effect sizes were replicable and comparable in a sample of 23 NIRS studies investigating infants' abilities to process repetition- and diversity-based regularities in linguistic and nonlinguistic auditory and visual sequences. Additionally, we tested whether effect sizes were modulated by different factors such as the age of participants or the laboratory. We obtained NIRS data from 12 published and 11 unpublished studies. The 23 studies involved a total of 487 infants, aged between 0 and 9 months, tested in four different countries (Canada, France, Italy, and USA).ResultsOur most important finding is that study and laboratory were never significant moderators of variation in effect sizes, indicating that results replicated reliably across the different studies and labs included in the sample. We observed small-to-moderate effect sizes, similar to effect sizes found with other neuroimaging and behavioral techniques in the developmental literature. In line with existing findings, effect sizes were modulated by the participants' age and differed across the different regularities tested, with repetition-based regularities giving rise to the strongest effects; in particular, the overall magnitude of this effect in the left temporal region was 0.27 when analyzing the entire dataset.ConclusionsMeta-analysis is a useful tool for assessing replicability and cross-study variability. Here, we have shown that infant NIRS studies in the language domain replicate robustly across various NIRS machines, testing sites, and developmental populations.

Dataset Information

Meta-analytic framework for sparse K-means to identify disease subtypes in multiple transcriptomic studies.

Publications

Meta-analytic framework for sparse <i>K</i>-means to identify disease subtypes in multiple transcriptomic studies.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets