Efficient feature selection and multiclass classification with integrated instance and model based learning

Predictive health monitoring: Leveraging artificial intelligence for early detection of infectious diseases in nursing home residents through discontinuous vital signs analysis

This research addresses the problem of detecting acute respiratory, urinary tract, and other infectious diseases in elderly nursing home residents using machine learning algorithms. The study analyzes data extracted from multiple vital signs and other contextual information for diagnostic purposes. The daily data collection process encounters sampling constraints due to weekends, holidays, shift changes, staff turnover, and equipment breakdowns, resulting in numerous nulls, repeated readings, outliers, and meaningless values. The short time series generated also pose a challenge to analysis, preventing the extraction of seasonal information or consistent trends. Blind data collection results in most of the data coming from periods when residents are healthy, resulting in excessively imbalanced data. This study proposes a data cleaning process and then builds a mechanism that reproduces the basal activity of the residents to improve the classification of the disease. The results show that the proposed basal module-assisted machine learning techniques allow anticipating diagnostics 2, 3 or 4 days before doctors decide to start treatment with antibiotics, achieving a performance measured by the area-under-the-curve metric of 0.857. The contributions of this work are: (1) a new data cleaning process; (2) the analysis of contextual information to improve data quality; (3) the generation of a baseline measure for relative comparison; and (4) the use of either binary (disease/no disease) or multiclass classification, differentiating among types of infections and showing the advantages of multiclass versus binary classification. From a medical point of view, the anticipated detection of infectious diseases in institutionalized individuals is brand new.

The diagnostic potential and barriers of microbiome based therapeutics

High throughput technological innovations in the past decade have accelerated research into the trillions of commensal microbes in the gut. The 'omics' technologies used for microbiome analysis are constantly evolving, and large-scale datasets are being produced. Despite of the fact that much of the research is still in its early stages, specific microbial signatures have been associated with the promotion of cancer, as well as other diseases such as inflammatory bowel disease, neurogenerative diareses etc. It has been also reported that the diversity of the gut microbiome influences the safety and efficacy of medicines. The availability and declining sequencing costs has rendered the employment of RNA-based diagnostics more common in the microbiome field necessitating improved data-analytical techniques so as to fully exploit all the resulting rich biological datasets, while accounting for their unique characteristics, such as their compositional nature as well their heterogeneity and sparsity. As a result, the gut microbiome is increasingly being demonstrating as an important component of personalised medicine since it not only plays a role in inter-individual variability in health and disease, but it also represents a potentially modifiable entity or feature that may be addressed by treatments in a personalised way. In this context, machine learning and artificial intelligence-based methods may be able to unveil new insights into biomedical analyses through the generation of models that may be used to predict category labels, and continuous values. Furthermore, diagnostic aspects will add value in the identification of the non invasive markers in the critical diseases like cancer.

Subtyping and grading of lower-grade gliomas using integrated feature selection and support vector machine

Classifying lower-grade gliomas (LGGs) is a crucial step for accurate therapeutic intervention. The histopathological classification of various subtypes of LGG, including astrocytoma, oligodendroglioma and oligoastrocytoma, suffers from intraobserver and interobserver variability leading to inaccurate classification and greater risk to patient health. We designed an efficient machine learning-based classification framework to diagnose LGG subtypes and grades using transcriptome data. First, we developed an integrated feature selection method based on correlation and support vector machine (SVM) recursive feature elimination. Then, implementation of the SVM classifier achieved superior accuracy compared with other machine learning frameworks. Most importantly, we found that the accuracy of subtype classification is always high (>90%) in a specific grade rather than in mixed grade (~80%) cancer. Differential co-expression analysis revealed higher heterogeneity in mixed grade cancer, resulting in reduced prediction accuracy. Our findings suggest that it is necessary to identify cancer grades and subtypes to attain a higher classification accuracy. Our six-class classification model efficiently predicts the grades and subtypes with an average accuracy of 91% (±0.02). Furthermore, we identify several predictive biomarkers using co-expression, gene set enrichment and survival analysis, indicating our framework is biologically interpretable and can potentially support the clinician.

A Deep Learning-Based Framework for Supporting Clinical Diagnosis of Glioblastoma Subtypes

Understanding molecular features that facilitate aggressive phenotypes in glioblastoma multiforme (GBM) remains a major clinical challenge. Accurate diagnosis of GBM subtypes, namely classical, proneural, and mesenchymal, and identification of specific molecular features are crucial for clinicians for systematic treatment. We develop a biologically interpretable and highly efficient deep learning framework based on a convolutional neural network for subtype identification. The classifiers were generated from high-throughput data of different molecular levels, i.e., transcriptome and methylome. Furthermore, an integrated subsystem of transcriptome and methylome data was also used to build the biologically relevant model. Our results show that deep learning model outperforms the traditional machine learning algorithms. Furthermore, to evaluate the biological and clinical applicability of the classification, we performed weighted gene correlation network analysis, gene set enrichment, and survival analysis of the feature genes. We identified the genotype-phenotype relationship of GBM subtypes and the subtype-specific predictive biomarkers for potential diagnosis and treatment.

LVQ-KNN: Composition-based DNA/RNA binning of short nucleotide sequences utilizing a prototype-based k-nearest neighbor approach

Unbiased sequencing is an upcoming method to gain information of the microbiome in a sample and for the detection of unrecognized pathogens. There are many software tools for a taxonomic classification of such metagenomics datasets available. Numerous of them have a satisfactory sensitivity and specificity for known organisms, but they fail if the sample contains unknown organisms, which cannot be detected by similarity-based classification employing available databases. However, recognition of unknowns is especially important for the detection of newly emerging pathogens, which are often RNA viruses. Here we present the composition-based analysis tool LVQ-KNN for binning unclassified nucleotide sequence reads into their provenance classes DNA or RNA. With a 5-fold cross-validation, LVQ-KNN reached correct classification rates (CCR) of up to 99.9% for the classification into DNA/RNA. Real datasets gained CCRs of up to 94.5%. Comparing the method to another composition-based analysis tool, similar or better classification results were reached. LVQ-KNN is a new tool for DNA/RNA classification of sequence reads from unbiased sequencing approaches that could be applicable for the detection of yet unknown RNA viruses in metagenomic samples. The source-code, training and test data for LVQ-KNN is available at Github (https://github.com/ab1989/LVQ-KNN).

Classification of metagenomics data at lower taxonomic level using a robust supervised classifier

As more and more completely sequenced genomes become available, the taxonomic classification of metagenomic data will benefit greatly from supervised classifiers that can be updated instantaneously in response to new genomes. Currently, some supervised classifiers have been developed to assess the organism of metagenomic sequences. We have found that the existing supervised classifiers usually cannot discriminate the training data from different classes accurately when the data contain some outliers. However, the training genomic data (bacterial and archaeal genomes) usually contain a portion of outliers, which come from sequencing errors, phage invasions, and some highly expressed genes, etc. The outliers, treated as noises, prohibit the development of classifiers with better prediction accuracy. To solve the problem, we present a robust supervised classifier, weighted support vector domain description (WSVDD), which can eliminate the interference from some outliers for training genomic data and then generate more accurate data domain descriptions for each taxonomic class. The experimental results demonstrate WSVDD is more robust than other classifiers for simulated Sanger and 454 reads with different outlier rates. In addition, in experiments performed on simulated metagenomes and real gut metagenomes, WSVDD also achieved better prediction accuracy than other classifiers.