Tensor decomposition based and principal component analysis based unsupervise...Y-h Taguchi
The document summarizes a presentation that was planned to be given at APBC2018 but was canceled. The presentation aimed to identify genes associated with differences in gene expression and methylation profiles between queens and workers in two social insect species (Dinoponera quadriceps and Polistes canadensis) using unsupervised feature extraction methods like principal component analysis and tensor decomposition. The analysis identified genes that were distinctly methylated and expressed between castes for each species, which were validated through gene ontology enrichment.
Tensor decomposition based unsupervised feature extraction applied to matrix...Y-h Taguchi
The document proposes using tensor decomposition on matrix products of multi-view data to perform unsupervised feature extraction. It applies this method to synthetic data containing embedded correlated variables and multi-omics data involving mRNA and miRNA. For the synthetic data, the method recovers the original orthogonal vectors after identifying correlations. For the multi-omics data, a few inter-correlated miRNAs and mRNAs are identified as statistically significant among many features.
Identification of Candidate Drugs for Heart Failure using Tensor Decompositio...Y-h Taguchi
This document describes a tensor decomposition-based unsupervised feature extraction method for identifying candidate drugs for heart failure using integrated analysis of gene expression data from heart failure patients and drugs. The method involves generating a tensor from the mathematical product of gene expression data between human diseases and drug-treated model animals. Tensor decomposition is applied to extract features without supervised training. The method successfully identified genes affected by drug treatments that could be biologically reasonable candidate targets. Further applications combining additional drug and target gene data are underway.
Tensor decomposition based and principal component analysis based unsupervise...Y-h Taguchi
The document summarizes a presentation that was planned to be given at APBC2018 but was canceled. The presentation aimed to identify genes associated with differences in gene expression and methylation profiles between queens and workers in two social insect species (Dinoponera quadriceps and Polistes canadensis) using unsupervised feature extraction methods like principal component analysis and tensor decomposition. The analysis identified genes that were distinctly methylated and expressed between castes for each species, which were validated through gene ontology enrichment.
Tensor decomposition based unsupervised feature extraction applied to matrix...Y-h Taguchi
The document proposes using tensor decomposition on matrix products of multi-view data to perform unsupervised feature extraction. It applies this method to synthetic data containing embedded correlated variables and multi-omics data involving mRNA and miRNA. For the synthetic data, the method recovers the original orthogonal vectors after identifying correlations. For the multi-omics data, a few inter-correlated miRNAs and mRNAs are identified as statistically significant among many features.
Identification of Candidate Drugs for Heart Failure using Tensor Decompositio...Y-h Taguchi
This document describes a tensor decomposition-based unsupervised feature extraction method for identifying candidate drugs for heart failure using integrated analysis of gene expression data from heart failure patients and drugs. The method involves generating a tensor from the mathematical product of gene expression data between human diseases and drug-treated model animals. Tensor decomposition is applied to extract features without supervised training. The method successfully identified genes affected by drug treatments that could be biologically reasonable candidate targets. Further applications combining additional drug and target gene data are underway.
microRNA-mRNA interaction identification in Wilms tumor using principal compo...Y-h Taguchi
This document describes a study that uses principal component analysis (PCA) based unsupervised feature extraction to identify microRNA-mRNA interactions in Wilms tumor. PCA is applied to gene expression data to select outlier mRNAs and miRNAs without using predetermined significance thresholds. Selected mRNA-miRNA pairs are validated using existing miRNA-mRNA interaction databases and survival analysis. The approach identifies more biologically feasible mRNAs than only considering differential gene expression between normal and tumor samples.
Principal component analysis-based unsupervised feature extraction applied to...Y-h Taguchi
1) The document presents two principal component analysis (PCA)-based unsupervised feature extraction methods (VBPCAFE and CPCAFE) for analyzing gene/microRNA expression data from stressed mouse hearts to understand posttraumatic stress disorder (PTSD)-mediated heart disease.
2) In synthetic tests, unsupervised methods outperformed supervised methods when there was label noise, demonstrating robustness to mislabeling.
3) When applied to real biological gene expression data, CPCAFE identified features with expected biological relationships and terms relevant to heart disease and neurodegeneration, performing better than supervised methods.
Comparison of Hepatocellular Carcinoma miRNA Expression Profiling as Evaluate...Y-h Taguchi
1) The study evaluated consistency between miRNA expression profiles of hepatocellular carcinoma (HCC) samples as measured by next generation sequencing (miRNA-seq) using MiSeq and miRDeep2, and microarray.
2) miRNA-seq and microarray results were highly correlated (r=0.6059) without complex normalization procedures. Differential expression between samples was also well correlated.
3) Eleven miRNAs were able to accurately diagnose HCC samples versus normal controls based on miRNA-seq data, demonstrating its ability to characterize HCC.
4) miRNA-seq using MiSeq and miRDeep2 produced reproducible results and could identify novel miRNAs.
microRNA-mRNA interaction identification in Wilms tumor using principal compo...Y-h Taguchi
This document describes a study that uses principal component analysis (PCA) based unsupervised feature extraction to identify microRNA-mRNA interactions in Wilms tumor. PCA is applied to gene expression data to select outlier mRNAs and miRNAs without using predetermined significance thresholds. Selected mRNA-miRNA pairs are validated using existing miRNA-mRNA interaction databases and survival analysis. The approach identifies more biologically feasible mRNAs than only considering differential gene expression between normal and tumor samples.
Principal component analysis-based unsupervised feature extraction applied to...Y-h Taguchi
1) The document presents two principal component analysis (PCA)-based unsupervised feature extraction methods (VBPCAFE and CPCAFE) for analyzing gene/microRNA expression data from stressed mouse hearts to understand posttraumatic stress disorder (PTSD)-mediated heart disease.
2) In synthetic tests, unsupervised methods outperformed supervised methods when there was label noise, demonstrating robustness to mislabeling.
3) When applied to real biological gene expression data, CPCAFE identified features with expected biological relationships and terms relevant to heart disease and neurodegeneration, performing better than supervised methods.
Comparison of Hepatocellular Carcinoma miRNA Expression Profiling as Evaluate...Y-h Taguchi
1) The study evaluated consistency between miRNA expression profiles of hepatocellular carcinoma (HCC) samples as measured by next generation sequencing (miRNA-seq) using MiSeq and miRDeep2, and microarray.
2) miRNA-seq and microarray results were highly correlated (r=0.6059) without complex normalization procedures. Differential expression between samples was also well correlated.
3) Eleven miRNAs were able to accurately diagnose HCC samples versus normal controls based on miRNA-seq data, demonstrating its ability to characterize HCC.
4) miRNA-seq using MiSeq and miRDeep2 produced reproducible results and could identify novel miRNAs.