Research

My main research interests and contributions are focused on devising new shallow and deep machine learning algorithms with applications in interactomics and transcriptomics. The main applications are on general problems in these fields and on the role of these in finding biomarkers in different type of cancer, as well as on the role of motifs and other predictive features in protein interactions, Calmodulin binding proteins and drug-protien interactions. Below is a summary of my interests and contributions in the respective fields.

Note: This page is not being updated frequently and some of the newer research projects are not listed here. For more details on my recent research contributions, refer to my List of Publications, or follow my research and publications in the most popular scholar databases:

Research Gate	Google Scholar	DBLP

Machine Learning

Computers can learn like human beings, from observations, examples, images, sensors, data, and experience. Machine learning is a field of artificial intelligence that involves the design and implementation of algorithms for computers to evolve their behavior from observations, examples, images, sensors, data, and experience, among many other sources. My main research in machine learning is on developing shallow and deep learning algorithms for classification, clustering, feature selection, dimensionality reduction, representation learning and performance evaluation, with applications to interactomics, transcriptomics and data integration. One of my most recent works is iSOM-GSN, a novel approach that combines self-organizing maps and convolutional neural networks for deriving gene similarity networks and prediction of disease states represented as graphs. More information about my research in this field can be found in the Machine Learning page.

Transcriptomics

The transcriptome represents the repertoire of transcripts in an organism as the main product of DNA transcription and splicing. The Human genome comprises 3 billion bases on each of (on average 10¹⁴) cells in one body, where each cell may contain up to 300k RNA molecules. Then, the full transcriptome may contain approximately 8.4²³ RNA bases... in one body! One cell line/condition, manifested in terms of transcriptomics data, could imply 30Mb of microarray data or 30Gb of RNA-seq data, while for all cells in one body the figure could grow up to petabytes or exabytes.

Transcriptomics studies have been traditionally carried out using microarray technologies and more recently, using the emerging next generation sequencing techniques known as RNA-seq. My main research in transcriptomics has been centered in the main aspects of microarray data analysis, with emphasis on DNA microarray image gridding and segmentation, as well as gene selection, biomarker detection and clustering time-time series gene expression data.

Currently, my research focuses on RNA-seq data analysis, aiming at finding relevant enriched regions in RNA-seq and ChIP-seq data and studying the underlying mechanisms of alternative splicing, its relationship with non-coding RNA, as well as their translation into protein isoforms and associated functions, with applications to the discovery of new biomarkers in breast and prostate cancers. One of the main goals is to understand the relationships among different protein variants yielded by alternative splicing and the integration with interactomics: the inherent functions as a result of protein interactions, the underlying domains and short motifs involved, and the dynamics of the interactome. A more recent project involves the study of spatial relationships among cells in single-cell RNA-seq data using graph representation learning approaches. More details are in the Transcriptomics page.

Interactomics

The interactome involves proteins and associated molecules interacting in a living system. The interactome is rather dynamic as the interactions and ultimately proteins' functions are manifested in a temporal and spatial manner. To understand the complex cellular mechanisms involved in a biological system, it is necessary to study the nature and specificity of these interactions and the dynamics involved in it at the molecular level, for which prediction of protein-protein interactions (PPIs) has played a significant role. In a broad sense, my main research in interactomics aims to develop machine learning algorithms for prediction and analysis of PPIs from high-throughput data, understanding the dynamic aspects of these interactions and their relationships with genomic and transcriptional features. One of the key issues I am currently investigating is the integration of transcriptomics data from RNA-seq with interactomics in applications for the identification of biomarkers that will help understand the transcriptional and genetic mechanisms involved in the development of prostate and breast cancer. Another aspect of protein interations that my lab is currently involved in is in calmoudlin-biding proteins and their interactions with other proteins. More information about my research in this field can be found in the Interactomics page.