Geoffrey Solano

Lung cancer classification using genetic algorithm to optimize prediction models

Lung cancer is one of the most fatal types of cancer around the world. The World Cancer Research Fund International estimated that in 2012, 1.8 million new cases of this disease were diagnosed. Early diagnosis and classification of this condition prompts medical professionals on safer and more effective treatment of the patient. Availability of microarray technology has paved the way to exploring the genes and its association in various diseases like lung cancer. This study utilized genetic algorithm as a method of feature (genes) selection for the support vector machine and artificial neural network to classify lung cancer status of a patient. Genetic algorithm (GA) successfully identified genes that classify patient lung cancer status with notable predictive performance.

Plant leaf recognition by venation and shape using artificial neural networks

The number of known and unknown plant species increases as time goes by. Research on plant species can be further advanced if there is a quick and accurate system that can identify plants and hasten the classification process. This system will not only help in accelerating plant classification, but will also allow people who are not morphological experts to conduct their own studies. LeaVes is an application designed to classify different plant species based on the leafs shape and venation. This system uses different image processing and machine learning techniques including centroid-radii, moment invariance, canny edge detection, morphological operations, image difference and artificial neural networks.

RadSS: A radiolarian classifier using support vector machines

Radiolarian assemblages have played a significant role as a biostratigraphic and paleoenvironmental tool used in the geological settings. These species can be used in studying sediments lacking calcareous fossils. Easy identification of these species would allow micropaleontologists to proceed further into studying the structure and way of living of these Radiolarians. RaDSS is a decision support system that could help researchers in classifying microphotographs of Radiolarian species through image processing and machine learning algorithms such as SVM.

Microarray data clustering and visualization tool using self-organizing maps

Microarray is one of the technologies used in the interdisciplinary science of Biolnformatics. Its primary objective is to discover biological knowledge among genes through their expressions. Gene expressions usually come in large and multidimensional data which makes computational and statistical analyses necessary. Clustering of microarray data is one of these. Grouping similar genes together unfolds relationships of the biological properties of the genes under specific condition and, if supported by visualization, serves as good decision support for researchers. MaSOM is a software that uses Self-Organizing Maps, an Artificial Neural Network suitable both for clustering and for visualization. This tool can be used to analyze large data set by preprocessing, clustering, and visualizing two-color cDNA microarray data. It can therefore aid microarray researchers and practitioners in determining the initial properties of the data they study before proceeding to their actual experimentation onto their data.

Lung cancer classification tool using microarray data and support vector machines

Lung cancer is one of the deadliest types of cancer around the world. Epidemiologic studies have shown that genetic variability is among the factors that affect a persons susceptibility to lung cancer. A recent study conducted by a team of researchers from the United States National Cancer Institute among 14,000 Asian women found out that Asian women, whether smokers or not, are more prone to developing cancer due to their genetic variations. This study proposes a system that utilizes gene expression data from oligonucleotide microarrays to predict the presence or absence of lung cancer, predict the specific type of lung cancer should it be present, and determine marker genes that are attributable to the specific kind of the disease. The proposed system would help in the faster diagnosis and serve as a reliable adjunct approach to current lung cancer classification methods.

Building phylogenetic trees from frequent subgraph mining techniques on reaction hypergraphs

Huge advances in experimental techniques have resulted in increasing amounts of biological network data being made available. Different kinds metabolic networks are just some of these. Analyzing the network topology of these metabolic networks across taxa can uncover important biological information that is independent of other currently available biological information. This study explores topological similarities between metabolic networks of different taxa to build phylogenetic trees. Similarities between graphs were determined using frequent subgraph mining techniques. Phylogenetic trees were then built based on the frequent subgraphs among reaction hypergraphs of different taxa. Experimental results show phylogenetic trees that bear a striking resemblance to those obtained from sequence comparisons.

Modelling the Coverage of Dipterocarp Trees in Central Visayas, Philippines

With the rapid decline of dipterocarp coverage in the Philippines, efforts in restoration would benefit from having a suitability map that would indicate the suitable areas where different dipterocarp species could thrive. Another use of these maps would be for assessing whether particular locations should be considered as Protected Areas in which exploitation would be limited, if not completely prohibited. We obtained data from the Department of Environment and Natural Resources of the Philippines, which consists of dipterocarp-related information gathered over localities in the countrys Central Visayas region. Six climate data parameters were then chosen and obtained from the Philippine Atmospheric Geophysical and Astronomical Services Administration. We employed a maximum entropy-based niche modelling approach to generate a suitability map. The model that we produced obtained an area under the curve (AUC) score of 0.955 with minimum temperature being the variable with the highest contribution. The map produced indicates that dipterocarp trees are more likely to appear in the lower regions of Central Visayas.