Yuval Kluger PhD

Associate Professor of Pathology

Research Interests

Signal processing and dimensional reduction of genome-wide data; Local and non-local genomic pattern recognition; Identification of cancer subclones with proliferation and invasion potential in heterogeneous cancer biopsies

Current Projects

• In silico de-mixing of genomics signals from heterogeneous tumor cell populations into their leading subclonal components (http://arxiv.org/abs/1301.1966)
• Biomarker discovery in whole-exome sequencing, RNA-Seq and immunohistochemistry
• Approaches for studying the epigenetic landscape at different length scales
• Combining 4C-seq, FISH and epigenetics for studying translocations in immune cells
• Making sense of diversity of mutation profiles within specific cancer populations
• Boosting software tool performance via effective validations and crowd-sourcing
• (http://arxiv.org/abs/1303.3257 and http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3192143/)
• Event detection in genomics data
• Nonlinear dimensional reduction approaches for detecting genomic patterns
• Inferring isoform usage by combining RNA-seq profiles from multiple states
• Building user-friendly frameworks for ChIP-Seq analysis

Research Summary

The research in our computational biology and bionformatics laboratory involves analysis of genomics and proteomics experiments. This includes computational analysis of output from high-throughput datasets generated from experiments involving melanoma, breast cancer, hematopoeisis, cell cycle genomics, and protein-protein interactions. The central focus of our earlier studies was to reveal functional and regulatory gene modules using genome-wide data generated in various "Omics" experiments and auxiliary information from genomics databases. We addressed issues of normalization and artifacts in microarrays. Subsequently, we developed a novel spectral method for bi-directional clustering of cancer microarray data to reveal regulatory gene modules. The lab has also focused on extracting meaningful biological information from experimental systems by assessing the co-expression of genes regulated by various transcription factors, evaluating pathway expression and building genetic networks based on functionality rather than pure expression. This approach is a step forward in identifying genes in regulatory networks that are disrupted by mutations of tumor suppressors and oncogenes and could shed light on the process of malignant transformation. Our research also involves the integration of sequence information with genome-wide transcriptome and epigenome profiles. This analysis has allowed us and our collaborators to reveal non-unique sequence recognition motifs of transcription factors in an in vivo context and to predict combinatorial regulation partners of transcription factors. Moreover, this approach has allowed us to find spatial organization of transcription factor binding events, as well as their relationships with other epigenomic factors.

The current computational activities in our laboratory include the following areas: a) Application of signal processing approaches for identification of relevant biological signals in high-throughput experiments, such as identification of aberrations in multi-subclonal cancer samples, signal denoising in next generation platforms, and de-mixing of cell types in heterogeneous samples, b) developing approaches to analyze high dimensional data from genomics platforms for biomarker discovery and personalized medicine. In particular, we use advanced applied mathematical methods to search complex local and non-local genomic patterns across the genome that may discriminate cancer patients with good vs. poor outcomes in CNA studies employing next generation sequencing or SNP platforms and c) uncovering direct and collective regulatory relationships between regulators (TFs, epigenomic factors and miRNAs) and their target genes by integration of heterogeneous Omics datasets and DNA sequences.

From a biological standpoint we are particularly interested in: a) Identification of primary or drug-treated metastatic subclones with proliferation and invasion potential in heterogeneous cancer biopsies b) The interplay between regulatory motifs, chromatin status and multi scale chromosomal structure c) Determining whether complex traits associated with certain common diseases vary across populations with different genetic backgrounds

Extensive Research Description

The research in our computational biology and bionformatics laboratory involves analysis of genomics and proteomics experiments. This includes computational analysis of output from high-throughput datasets generated from experiments involving melanoma, breast cancer, hematopoeisis, cell cycle genomics, and protein-protein interactions. The central focus of our earlier studies was to reveal functional and regulatory gene modules using genome-wide data generated in various "Omics" experiments and auxiliary information from genomics databases. We addressed issues of normalization and artifacts in microarrays. Subsequently, we developed a novel spectral method for bi-directional clustering of cancer microarray data to reveal regulatory gene modules. The lab has also focused on extracting meaningful biological information from experimental systems by assessing the co-expression of genes regulated by various transcription factors, evaluating pathway expression and building genetic networks based on functionality rather than pure expression. This approach is a step forward in identifying genes in regulatory networks that are disrupted by mutations of tumor suppressors and oncogenes and could shed light on the process of malignant transformation. Our research also involves the integration of sequence information with genome-wide transcriptome and epigenome profiles. This analysis has allowed us and our collaborators to reveal non-unique sequence recognition motifs of transcription factors in an in vivo context and to predict combinatorial regulation partners of transcription factors. Moreover, this approach has allowed us to find spatial organization of transcription factor binding events, as well as their relationships with other epigenomic factors.

The current computational activities in our laboratory include the following areas: a) Application of signal processing approaches for identification of relevant biological signals in high-throughput experiments, such as identification of aberrations in multi-subclonal cancer samples, signal denoising in next generation platforms, and de-mixing of cell types in heterogeneous samples, b) developing approaches to analyze high dimensional data from genomics platforms for biomarker discovery and personalized medicine. In particular, we use advanced applied mathematical methods to search complex local and non-local genomic patterns across the genome that may discriminate cancer patients with good vs. poor outcomes in CNA studies employing next generation sequencing or SNP platforms and c) uncovering direct and collective regulatory relationships between regulators (TFs, epigenomic factors and miRNAs) and their target genes by integration of heterogeneous Omics datasets and DNA sequences.

From a biological standpoint we are particularly interested in: a) Identification of primary or drug-treated metastatic subclones with proliferation and invasion potential in heterogeneous cancer biopsies b) The interplay between regulatory motifs, chromatin status and multi scale chromosomal structure c) Determining whether complex traits associated with certain common diseases vary across populations with different genetic backgrounds

Selected Publications