Analyzing time-dependent microarray data using independent component analysis derived expression modes from human macrophages infected with F. tularensis holartica.

The analysis of large-scale gene expression profiles is still a demanding and extensive task. Modern machine learning and data mining techniques developed in linear algebra, like Independent Component Analysis (ICA), become increasingly popular as appropriate tools for analyzing microarray data. We applied ICA to analyze kinetic gene expression profiles of human monocyte derived macrophages (MDM) from three different donors infected with Francisella tularensis holartica and compared them to more classical methods like hierarchical clustering. Results were compared using a pathway analysis tool, based on the Gene Ontology and the MeSH database. We could show that both methods lead to time-dependent gene regulatory patterns which fit well to known TNFalpha induced immune responses. In comparison, the nonexclusive attribute of ICA results in a more detailed view and a higher resolution in time dependent behavior of the immune response genes. Additionally, we identified NFkappaB as one of the main regulatory genes during response to F. tularensis infection.

Knowledge-based gene expression classification via matrix factorization.

MOTIVATION: Modern machine learning methods based on matrix decomposition techniques, like independent component analysis (ICA) or non-negative matrix factorization (NMF), provide new and efficient analysis tools which are currently explored to analyze gene expression profiles. These exploratory feature extraction techniques yield expression modes (ICA) or metagenes (NMF). These extracted features are considered indicative of underlying regulatory processes. They can as well be applied to the classification of gene expression datasets by grouping samples into different categories for diagnostic purposes or group genes into functional categories for further investigation of related metabolic pathways and regulatory networks. RESULTS: In this study we focus on unsupervised matrix factorization techniques and apply ICA and sparse NMF to microarray datasets. The latter monitor the gene expression levels of human peripheral blood cells during differentiation from monocytes to macrophages. We show that these tools are able to identify relevant signatures in the deduced component matrices and extract informative sets of marker genes from these gene expression profiles. The methods rely on the joint discriminative power of a set of marker genes rather than on single marker genes. With these sets of marker genes, corroborated by leave-one-out or random forest cross-validation, the datasets could easily be classified into related diagnostic categories. The latter correspond to either monocytes versus macrophages or healthy vs Niemann Pick C disease patients.

Comparison of unsupervised and supervised gene selection methods.

Modern machine learning methods based on matrix decomposition techniques like Independent Component Analysis (ICA) provide new and efficient analysis tools which are currently explored to analyze gene expression profiles. These exploratory feature extraction techniques yield informative expression modes (ICA) which are considered indicative of underlying regulatory processes. Their most strongly expressed genes represent marker genes for classification of the tissue samples under investigation. Comparison with supervised gene selection methods based on statistical scores or support vector machines corroborate these findings. The method is applied to macrophages loaded/de-loaded with chemically modified low density lipids.

How to extract marker genes from microarray data sets.

In this study we focus on classification tasks and apply matrix factorization techniques like principal component analysis (PCA), independent component analysis (ICA) and non-negative matrix factorization (NMF) to a microarray data set. The latter monitors the gene expression levels (GEL) of mononcytes and macrophages during and after differentiation. We show that these tools are able to identify relevant signatures in the deduced matrices and extract marker genes from these gene expression profiles (GEPs) without the need for extensive data bank search for appropriate functional annotations. With these marker genes corresponding test data sets can then easily be classified into related diagnostic categories.

Routes to identify marker genes for microarray classification.

Support vector machines are applied to extract marker genes from various microarray data sets: Breast Cancer, Leukemia and Monocyte - Macrophage Differentiation to ease classification of related pathologies or characterize related gene regulation pathways.

Diagnosis of nontraumatic hematobilia by computerized tomography of the abdomen.

Hematobilia was diagnosed by computerized tomography as the cause of gastrointestinal bleeding in a 73-yr-old man. The patient suffered from recurrent abdominal pain for 4 yr and presented with a history of episodic hematochezia over the previous 24 h, suggestive of significant lower gastrointestinal bleeding. Endoscopic evaluation of the lower and upper gastrointestinal tract revealed only blood in the region of a normal ampulla. Computerized tomography demonstrated a thickened gallbladder filled with material the density of blood. A blood-filled gallbladder was removed at cholecystectomy and pathological examination of the excised gallbladder demonstrated ulceration and inflammation of the mucosa. The diagnostic evaluation of hematobilia is discussed.