We are glad to invite you to a seminar by Prof. Dmitri Papatsenko, titled “Feedback Control of Self-Renewal and Memory of Pluripotent Stem Cells”

Seminar abstract

Embryonic stem cells (ESC) can remain in the pluripotent state and propagate or differentiate towards a certain cell lineage. It remains largely unknown how the ESCs maintain the meta-stable self-renewing pluripotency states and why ESC may eventually commit differentiation.

Based on single cell data and previously published data we have reconstructed pluripotency gene regulatory network (PGRN), and suggested a model for dynamic exchange between two identified pluripotency states in mouse ESCs. The dynamic exchange between the states suggested a mechanism of self-renewal, involving feedback control loops present in the PGRN.

During reprogramming from somatic cells, the resulting induced pluripotent stem cells (iPSC) can “remember” their original source cell type, as well as the reprogramming method or protocol.

Analysis of PGRN at many levels revealed multiple feedback control loops, suggesting that the PGRN architecture may resemble an architecture of a recurrent neural network (RNN). RNN is an artificial network used to model human memory. In the case of ESCs and iPSCs, the RNN-like network architecture may explain memory properties of the cells, emerging from reprogramming studies.

Speaker introduction

Prof. Papatsenko received his diploma (MS) in Molecular Biology from Moscow State University in 1991 and his PhD from Engelgardt Institute of Molecular Biology, Moscow in 1995.

In 1996 he moved to The Rockefeller University (New York, USA) to study gene regulation and development in Drosophila. At the Rockefeller University Prof. Papatsenko introduced systems analysis of gene control regions, promoters ad enhancers and proposed a number of models explaining formation of spatial gene expression patterns in early embryo of Drosophila.

In 2004 Prof. Papatsenko accepted position of a scientist at The University of California, Berkeley, where he developed an integrated  quantitative model explaining progressive spatial patterning of fly embryo from maternal gradients (determinants) to zygotic stripe expression patterns.

In 2010 Prof. Papatsenko moved to Icahn School of Medicine at Mount Sinai, New York, as an Assistant Professor to study gene networks in embryonic and hematopoietic stem cells. At Mount Sinai he worked on reconstruction of gene networks maintaining pluripotency and quantitative models explaining function of the pluripotency gene networks.

We are glad to invite you to a seminar by Dr. Dmitri Pervouchine, titled “Challenges of Big Data in Understanding RNA Biogenesis and Function”

Seminar abstract

High-throughput sequencing technology dramatically enhanced our capability to look at the molecular parts of living cells. Genomes, transcriptomes, chromatin, transcription factor binding, cellular localization — this is a largely incomplete list of objects and events that can now be studied at a previously unprecedented level of detail. Since the technology is still labor-intensive and costly, many research groups coalesced into big consortia in order to generate comprehensive compendia of data sets, each with a particular biological focus.

Quite a typical goal of a consortium work is to identify novel genomic elements or to quantify their abundance and associations with each other. This talk will be focused mainly on the data generated by human and mouse ENCODE projects, GTEx Genotype-Tissue expression project, and GENCODE reference gene annotations. I will explain seemingly conflicting results that have been published regarding whether organ specific transcriptional patterns should dominate over species specific patterns or vice versa. Next, I will describe IPSA, a pipeline for splicing analysis that was used to identify brain-specific micro-alterations of splicing. Besides consortium work I will present a few hypotheses on the biogenesis of circular RNAs, on the role of long-range RNA-RNA interactions in the regulation of splicing and polyadenylation, and share some thoughts about prospective projects on the interface between genomics and neuroscience.

The consortium format has been increasingly criticized for its its inability to produce integrated data sets and for its higher cost compared to small scale projects. In the light of lessons learned from working in large collaborative environments, I will share my view of the current challenges in management and integration of big data, perspectives on where the technology will next develop, and discuss what we can do to survive in the anticipated in silico data tsunami.

Speaker introduction

Dmitri Pervouchine was born in 1974 in Moscow, Russia. He graduated from Moscow State University with two degrees, in Mathematics and in Chemistry, and continued his education at the Bioinformatics program at Boston University. In 2002 he defended his PhD in Mathematics (Algebra) under the supervision of Dr. Ernest B. Vinberg and joined the group of Dr. Nancy Kopell at the Center for BioDynamics and Department of Mathematics and Statistics at Boston University, working on neuronal dynamics, memory, and learning. In 2005 he joined the group of Drs. Mikhail Gelfand and Andrei Mironov, where he worked on computational RNA structure prediction and evolution. Dmitri found a class of RNA structures in introns of eukaryotic genes that impact splicing of pre-mRNAs. In 2011 Dmitri joined transcriptomics research at the group of Dr. Roderic Guigo in the Center for Genomic Regulation in Barcelona. Dmitri works in close collaboration with Dr. Thomas Gingeras from Cold Spring Harbor Laboratory and particularly he was involved in mouse ENCODE consortium where he and colleagues identified a group of genes with constrained expression across mammalian evolution and described distinct properties of this group of genes.

We are glad to invite you to a seminar by Dr. Yuri Nikolsky, titled “OMICs Biomarkers and Drug Repurposing Via Knowledge-based Functional Analysis of Big Data”

Seminar abstract

Over the last two decades, Bid Data assays became the mainstream experimental technology in life sciences research and biomedicine. From mass market consumer genotyping at 23andme or Atlas, to elaborate multi-OMICs patient testing in clinical trials run by major drug companies, these assays generate multi-factorial “signatures” correlating with disease predisposition and severity, drug sensitivity, survival rates and many other important phenotypes. The main issue in OMICs experimentation is data interpretation: how to select the most relevant DNA variants, overexpressed genes, blood proteins from thousands of data points in sequencing files, microarrays, proteomics and other OMICs profiles?

In general, there are two somewhat complementary approaches to the problem: the “data-driven” and “knowledge-based” ones. The data-driven approach is based purely on statistical analysis and modeling; it selects the most mathematically relevant OMICs signals with complete ignorance of the biological functions of selected genes and variants. The second approach consists of capturing and annotation of  biological data from vast experimental literature, structuring it as protein interactions and pathways and applying this knowledge to the analysis of OMICs assays with help of mathematical algorithms of functional, or systems, analysis.

Over the years, our company GeneGo (acquired by Thomson Reuters) has created “from scratch” one of the most advanced platforms for functional data analysis, with the original semantics, annotation technology and algorithms. They have built a sophisticated database of over 1,000,000 protein-protein and protein-compound interactions, normal and pathological biological pathways, disease- and toxicity-related variants. Dr. Nikolsky will briefly describe the technology and show specific examples of applying this technology for predicting drug response, disease progression and drug repurposing.

Speaker introduction

Dr. Yuri Nikolsky is the Director of Science at the BMT cluster at Skolkovo Foundation. He is also an adjunct faculty at George Mason University in Fairfax, VA, USA. Yuri received his MS in genetics from Lomonosov Moscow University and Ph.D. in molecular biology from NII Genetika (Moscow). He also holds an MBA degree in finance from the Booth School of Business at the University of Chicago. Dr. Nikolsky is known internationally both as a scientist and a life sciences entrepreneur. After several years of research in meiosis at the University of Chicago, he co-founded Integrated Genomics (Chicago, IL), a pioneer sequencing and comparative genomics company. Later, he was CEO at two successful startups: ChemDiv (San Diego, CA) and GeneGo (San Diego, CA). After acquisition of GeneGo by a global informatics company Thomson Reuters, Yuri was employed as Vice President, R&D and built a service group focused on scientific consulting to the major pharmaceutical companies. Yuri’s current research interests include analysis of biological pathways activation and mining the data networks of biomedicine and life science finance.

We are glad to invite you to a seminar by Dr. Vsevolod Makeev, titled “Bioinformatics of Non-Coding Elements Responsible for Cell Identity Formation and Cancer”

When: January 28, 2016, 11.30 – 13.00
Where: TPOC-3, Room 148

Seminar abstract

One of the surprises of genome sequencing projects is the number of genes in multicellular organisms, which proved to be much smaller than expected and poorly correlated with organism “complexity”.  Particularly, the human lineage is a general gene loser through at least last 70 millions of years.  With human gene arsenal of less than 20.000, which is only 4-fold greater than the number of genes in some bacteria, human genome can determine one of ~200 types of each of about 10 trillion cells in the human body. This coding efficiency is explained by a complex regulatory elements controlling gene expression in each of the cell of a multicellular organism.   Cell identity formation involves the orchestrated activity of a plethora of genes at different stages of development.  To an extent sequence analysis can identify common DNA motifs responsible for controlled promoter activity. Although most of the human cells have roughly identical genomes, mutations happen as cells divide; such mutations can disrupt gene regulation.  Mutation density is tissue specific, and probably controlled by epigenetic modifications.      In my talk I will focus upon how statistical machine learning approaches applied to deep-sequencing data can unravel gene regulatory circuits controlling cell identity formation. I outline challenges in algorithm and database development still needed to harness the results of recent large-scale gene regulation projects.  Additionally, I will discuss how regulatory motifs withstand the omnipresent pressure of somatic mutations and what happens when a tissue loses its identity at cancer transformation.

Speaker introduction

Vsevolod Makeev is a leading scientist in the field of analysis of next generation sequencing data, particularly in the field of transcription regulation.  His main results were in development of algorithms for sequence analysis and databases of sequence motifs. Among other resources and tools he developed a Bayesian sequence segmenter Basio; ChIPmunk motif finder, one of the most powerful tools for ChIP-seq data analysis; and two databases of DNA motifs recognized by transcription factors: iDMMPMM for Drosophila melanogaster, HOCOMOCO for mouse and human. He took part in the analysis of human promoter tissue activity data as a member of FANTOM5 international consortium. Prof. Makeev teaches a course in Advanced Algorithms of Bioinformatics in Moscow Institute of Physics and Technology. He currently works in Vavilov Institute of General Genetics, Russian Academy of Sciences.