People - Zohar Yakhini

Laboratory of Computational Biology

Zohar Yakhini

who

Research Interests

Analysis of gene expression data. The routine use of microarrays to simultaneously measure the expression levels of thousands of genes across a variety of conditions and environments gives rise to a rich body of data, of a variety of types. Analyzing this data provides answers to a variety of biological questions, from the understanding of basic cellular mechanisms to the development of diagnostic marker sets. An overview of methods can be found in "Analysis of expression patterns: the scope of the problem, the problem of scope", by Chen et al.

Examples of statistical and algorithmic approaches and some applications are:

Clustering. Also please see an animation of CAST (by Benjamin Fry, MIT).
Analysis of classified expression data.
Class discovery. Agilent technical report, PNAS article, NHGRI website
Order preserving sub-matrices. See Ben-Dor et al’s work in the Proceedings of RECOMB 2002.

Analysis of SNP data. In collaboration with Amir Ben-Dor, Anya Tsalenko and others, we have developed methods that utilize information theory concepts and associated statistics to identify sets of SNPs that jointly explain trait differences. Some of these methods are described in our PSB 2003 paper.
Combinatorial optimization with applications to hybridization assay design. The design of efficient hybridization assays is key to make the information obtained in performing the assays biologically relevant and accurate. Specificity of hybridization probes determines the relevance of the resulting measurement signals. Highly specific probes enable the use of just a few of these to measure levels of target mRNAs. Statistical and algorithmic aspects of the design process are discussed in: Designing Specific Oligonucleotide Probes for the Entire S. cerevisiae Transcriptome. In universal array applications a code-design problem arises as described in Universal DNA Tag Systems: A Combinatorial Scheme Design.
Combinatorial optimization and novel measurement technologies. Combinatorial design questions that determine the efficiency of other measurement assays. In Optimally Multiplexed SNP Genotyping by Primer Extension and Mass-Spectrometry we study aspects of utilizing mass-spectrometry in genotyping assays. Together with Doron Lipson we are applying graph theory methods to optimize multiplex PCR design schemes.

A more inclusive list of publications can be found at the Computational Biology Lab main site.

Teaching

Students

Chaya Ben-Zaken Zilberstein, working on learning gene regulatory mecahnisms from expression and sequence data.
Ari Frank. Graduated with a MSc; co-advisor, jointly with Dan Geiger. Our research focused on feature selection for classified expression data.
Doron Lipson. Working on combinatorial optimization, especially in the context of hybridization assay design and on aspects of analyzing expression and other genomic data.
Eran Eden. Discovering Motifs in Ranked Lists of DNA Sequences. Thesis

Elinor Dehan. A MSc student of Naftali Kaminski. I act as a consultant to her research on common chromosomal aberrations in lung cancer.
Nurit Gal. A former student of Naftali Kaminski. I acted as a consultant to her research on the gene expression characteristics of PF fibroblasts and the effect of TGFbeta.
Efrat Manisterski. A student of Yonatan Aumann. Graduated with a MSc. I acted as a consultant to her research on optimizing SNP genotyping by mass-spectrometry.

Software

ATRHunter - find approximate tandem repeats in a genomic sequence. Developed by Ydo Wexler.
GeneSelector - find a small subset of genes for classification of expression data. Developed by Ari Frank.
ProbeSpec - for mapping the specificity of all candidate probes for a given sequence. Developed by Doron Lipson.

Contact Information

e-mail (Best!!!): zohar_yakhini@agilent.com
Office at the Technion: +972 4 829 4937
Cell-phone: +972 64 299 728