The publications on this page reflect my scientific work in the domains of grid computing, bioinfomatics, and molecular biology. They are sorted by type (research articles, review articles, book chapters, patents, interviews & general interest) and have brief comments attached.

Research articles:

Podvinec M, Lim SP, Schmidt T, Scarsi M, Wen D, Sonntag LS, Sanschagrin P, Shenkin PS, Schwede T: Novel inhibitors of dengue virus methyltransferase: discovery by in vitro-driven virtual screening on a desktop computer grid. J Med Chem. 2010 Feb 25;53(4):1483-95. [pdf]

In this paper, we describe the results of a combined virtual screening and laboratory assay study, targeting the Dengue virus methyltransferase. The aim of this study was to come up with novel small-molecule inhibitors of this enzyme, which could potentially be developed as drugs against Dengue fever.
We set out by using state-of-the-art approaches to virtual screening on a distributed computing infrastructure to screen over 5 million purchasable compounds for their potential to inhibit this enzyme. We then followed up these studies with more refined simulations and calculations, before ordering a selection of about 250 compounds, and assaying their ability to block this enzyme in laboratory assays. Therefore, this study has two focuses: On one hand, we present a number of new inhibitors of the dengue methyltransferase, on the other, we were able to retrospectively evaluate the individual steps of our in silico strategy, knowing the in vitro results.

Arnold K, Kiefer F, Kopp J, Battey JN, Podvinec M, Westbrook JD, Berman HM, Bordoli L, Schwede T: The Protein Model Portal. J Struct Funct Genomics. 2009 Mar;10(1):1-8. [pdf]

- and -

Berman HM, Westbrook JD, Gabanyi MJ, Tao W, Shah R, Kouranov A, Schwede T, Arnold K, Kiefer F, Bordoli L, Kopp J, Podvinec M, Adams PD, Carter LG, Minor W, Nair R, La Baer J: The protein structure initiative structural genomics knowledgebase. Nucleic Acids Res. 2009 Jan;37(Database issue):D365-8. [pdf]

The Protein Model Portal is a new single point of web-based access that gives users a unified, simple access to all model information leveraged from Protein Structure Initiative targets and other solved structures. Over 7 million protein models are presented in a consistent way and are mapped in a single reference space, where they can be compared to each other.

In the context of structural genomics, where structural information is determined for a large number of new protein structures, it is especially important to provide a clear overview of the homology models created with the help of these new structures. Often, structure models for hundreds of proteins can be derived from a single new experimentally determined structure, using a variety of established methods. The big challenge is how to use this model information effectively: Data for these models is stored in heterogeneous formats using incompatible accession codes and reference spaces.  The Protein Model Portal provides unified access to all these resources.

My contribution to this project was project management and coaching. I introduced the team to Scrum as a simple but highly efficient method to achieve project goals in a rapidly changing environment. This worked very well – we were able to deliver the first fully working, publicly released version of the Protein Model Portal after only six weeks worth of iterations. Since the initial release, the scope of the portal has continued to grow. Recently, the portal was integrated into Nature Publishing Group’s Structural Genomics Knowledgebase.

Moehren U, Denayer S, Podvinec M, Verrijdt G, Claessens: Identification of androgen-selective DNA response elements in the human Aquaporin 5 and RAD9 genes. Biochemical Journal (2008), 411(3):679-86. [pdf]

This publication is the outcome of a collaboration I started a couple of years ago with a Belgian laboratory. They contacted me after the publication of NUBIScan, requesting help in discovering further selective androgen-receptor elements in likely target genes. Based on the NUBIScan algorithm, I wrote some code that allowed the search for these special sites… Then, I didn’t hear back from them for a very long time. But it turns out that was because they were busy validating those predictions. In the end, we find ourselves writing a paper about it, a nice outcome.

Zosso D, Arnold K, Schwede T, Podvinec M: SWISS-TANDEM: A Web-based Workspace for MS/MS Protein Identification on PC Grid. In: Proceedings of the 20th IEEE International Symposium on Computer-Based Medical Systems (2007) [pdf]

This paper describes the SWISS-TANDEM web-based workspace for the analysis of protein mass spectrometry data. It is closely linked with the preceding paper [Zosso,HealthGrid2007], which describes the grid parallelization scheme underlying the workspace.

Zosso D, Podvinec M, Mueller M, Aebersold R, Peitsch MC, Schwede T: Tandem Mass Spectrometry Identification on a PC Grid. In: Proceedings of HealthGrid 2007, IOS Press, Amsterdam (2007). [pdf]

Analyzing protein mass spectrometry data is a computationally intense field, with refined protein search strategies suffering from combinatorial explosion.In this paper, we have investigated several parallelization strategies to address the problem on a grid architecture, and investigate their impact both on run-time and on the detection characteristics of the search.

Podvinec M, Scarsi M, Roth A, Hug H, Kersten S, Albrecht H, Schwede T, Meyer UA, Ruecker C: Sulfonylureas and Glinides Exhibit PPARgamma Activity: A Combined Virtual Screening and Biological Assay Approach Mol Pharmacol. 2007 Feb;71(2):398-406. [pdf]

Activating PPARγ causes insulin sensitizing in peripheral tissue and is a therapeutic strategy against type-2 diabetes mellitus. This study used high-throughput docking to discover new potential PPARγ ligands among known drugs and bioactive compounds. To our suprise, a number of the compounds ranked well in these computer simulations belong to the chemical classes of sulfonylureas and glinides, which both are used clinically against diabetes, but through a wholly different pathway, namely by stimulating insulin secretion. Intrigued, we followed up the simulations by biochemical and biological assays, and could indeed show binding, receptor activation, and activation of bona fide PPARγ target genes by these compounds

Podvinec M, Maffioletti S, Kunszt P, Arnold K, Cerutti L, Nyffeler B, Schlapbach R, Turker C, Stockinger H, Thomas AJ, Peitsch MC, Schwede T: The SwissBioGrid Project: Objectives, Preliminary Results and Lessons Learned In: 2nd IEEE International Conference on e-Science and Grid Computing (e-Science 2006), IEEE Computer Society Press. [pdf]

Through my interest in virtual screening, I became involved very early on with the SwissBioGrid initiative (http://www.swissbiogrid.org – now defunct). The first implementation of the SwissBioGrid middleware layer was done using ProtoGRID, a middleware developed by me and Konstantin Arnold, which is discussed here. The SwissBioGrid and its middleware aim to fulfil the specific requirements that biology-oriented grid computing has, and some of the lessons learned in the build-up process are discussed here.

Gnerre C, Schuster GU, Roth A, Handschin C, Johansson L, Looser R, Parini P, Podvinec M, Robertsson K, Gustafsson JA, Meyer UA: LXR deficiency and cholesterol feeding affect the expression and phenobarbital-mediated induction of cytochromes P450 in mouse liver. J Lipid Res. 2005 Aug;46(8):1633-42. [pdf]

This study investigated the interplay between members of the nuclear receptor families, especially the xenosensors and the LXRs as oxysterol sensors.

Handschin C, Blattler S, Roth A, Looser R, Oscarson M, Kaufmann MR, Podvinec M, Gnerre C, Meyer UA.: The evolution of drug-activated nuclear receptors: one ancestral gene diverged into two xenosensor genes in mammals. Nucl Recept. 2004 Oct 12;2(1):7. [pdf]

This paper follows up on our earlier work, and characterizes the chicken xenobiotic sensor (CXR) in more detail, in particular in relation to the two xenosensors (CAR and PXR) observed in mammals.

Podvinec M, Handschin C, Looser R, Meyer UA.: Identification of the xenosensors regulating human 5-aminolevulinate synthase. Proc Natl Acad Sci U S A. 2004 Jun 15;101(24):9127-32 [pdf]

This paper was the publication of one of the main threads of my Ph.D. thesis. I had started by dissecting the human ALAS1 gene, trying to figure out how xenobiotics directly induce its transcription, but not really succeeding. Four years later, and after a number of other projects, I came back to this initial question from a fresh perspective, and this time around, I was able to find out what makes ALAS1 respond to drugs.

Schreiber SN, Emter R, Hock MB, Knutti D, Cardenas J, Podvinec M, Oakeley EJ, Kralli A.: The estrogen-related receptor alpha (ERRalpha) functions in PPARgamma coactivator 1alpha (PGC-1alpha)-induced mitochondrial biogenesis. Proc Natl Acad Sci U S A. 2004 Apr 27;101(17):6472-7. [pdf]

I collaborated in this study and provided the computational analyses that successfully predicted the ERR alpha transcription factor binding sites.

Sanyal S, Handschin C, Podvinec M, Song KH, Kim HJ, Kim JY, Seo YW, Kim SA, Kwon HB, Lee K, Kim WS, Meyer UA, Choi HS.: Molecular cloning and characterization of chicken orphan nuclear receptor cTR2. Gen Comp Endocrinol. 2003 Jul;132(3):474-84. [pdf]

Meyer RP, Podvinec M, Meyer UA.: Cytochrome P450 CYP1A1 accumulates in the cytosol of kidney and brain and is activated by heme. Mol Pharmacol. 2002 Nov;62(5):1061-7. [pdf]

Fraser DJ, Podvinec M, Kaufmann MR, Meyer UA.: Drugs mediate the transcriptional activation of the 5-aminolevulinic acid synthase (ALAS1) gene via the chicken xenobiotic-sensing nuclear receptor (CXR). J Biol Chem. 2002 Sep 20;277(38):34717-26. [pdf]

This paper describes how CXR is able to transactivate the ALAS1 gene in chicken LMH cells in response to xenobiotic compounds. A first clear indication of the link between xenosensors and ALAS1.

Jung D, Podvinec M, Meyer UA, Mangelsdorf DJ, Fried M, Meier PJ, Kullak-Ublick GA.: Human organic anion transporting polypeptide 8 promoter is transactivated by the farnesoid X receptor/bile acid receptor. Gastroenterology. 2002 Jun;122(7):1954-66. [pdf]

I contributed promoter analysis to this study, which was crucial in finding out the way that FXR/BAR upregulates OATP8.

Handschin C, Podvinec M, Amherd R, Looser R, Ourlin JC, Meyer UA.: Cholesterol and bile acids regulate xenosensor signaling in drug-mediated induction of cytochromes P450. J Biol Chem. 2002 Aug 16;277(33):29561-7. [pdf]

This work showed that not only PXR and CAR, but also the nuclear receptors sensing levels of oxysterols and bile acids, FXR and LXR, play a role in drug induction, by modulating the effect of drugs on CYP450 induction.

Podvinec M, Kaufmann MR, Handschin C, Meyer UA.: NUBIScan, an in silico approach for prediction of nuclear receptor response elements. Mol Endocrinol. 2002 Jun;16(6):1269-79. [pdf]

This work describes an algorithm to predict several types of nuclear receptor response elements, using a positional weight-matrix approach, and demonstrates the usefulness of the proposed approach in a study of the chicken CYP3A37 gene. The algorithm is available at http://www.nubiscan.unibas.ch

Handschin C, Podvinec M, Looser R, Amherd R, Meyer UA.: Multiple enhancer units mediate drug induction of CYP2H1 by xenobiotic-sensing orphan nuclear receptor chicken xenobiotic receptor. Mol Pharmacol. 2001 Oct;60(4):681-9. [pdf]

A detailed dissection of the 5′ flanking region of the CYP2H1 gene, describing the drug response elements contained within.

Handschin C, Podvinec M, Stockli J, Hoffmann K, Meyer UA.: Conservation of signaling pathways of xenobiotic-sensing orphan nuclear receptors, chicken xenobiotic receptor, constitutive androstane receptor, and pregnane X receptor, from birds to humans. Mol Endocrinol. 2001 Sep;15(9):1571-85. [pdf]

This paper looks at CXR from an evolutionary standpoint, and shows a number of experiments that examine to what extent the components of the xenobiotics sensing machinery are interchangeable between species.

Handschin C, Podvinec M, Meyer UA.: CXR, a chicken xenobiotic-sensing orphan nuclear receptor, is related to both mammalian pregnane X receptor (PXR) and constitutive androstane receptor (CAR). Proc Natl Acad Sci U S A. 2000 Sep 26;97(20):10769-74. [pdf]

In this paper, we presented and characterized the chicken xenobiotic-sensing receptor (CXR) and its role in the induction of CYP2H1 and other drug-metabolizing enzymes.

Review articles:

Podvinec M, Meyer UA: Prediction of cis-regulatory elements for drug-activated transcription factors in the regulation of drug-metabolizing enzymes and drug transporters. Expert Opin. Drug Metab. Toxicol. (2006) 2(3) 367-79 [pdf]

This article was inspired by a short course I taught at the ISSX meeting in Vancouver. It is a tour of the current state of affairs in discovering regulatory pathways computationally, written with a biologist/pharmacologist audience in mind.

Handschin C, Podvinec M, Meyer UA.: In silico approaches, and in vitro and in vivo experiments to predict induction of drug metabolism. Drug News Perspect. 2003 Sep;16(7):423-34. [pdf]

Book chapters:

Podvinec M, Schwede T and Peitsch MC: Docking for Neglected Diseases as Community Efforts. In: Computational Structural Biology: Methods and Applications, World Scientific Publishing Co. (2008)

This was my first experience writing for a book. Writing a paper these days, you take into account the half-life of the information you’re about to present in a rapidly evolving field. Knowing that your words will ultimately be printed and bound into a nice hardcover volume, you adopt a much longer perspective. I totally underestimated the time this would take, though.

Patents:

US Patent US2005118601: MEYER URS (CH); FRASER DAVID J (CH); KAUFMANN MICHEL R (CH);PODVINEC MICHAEL (CH); ZUMSTEG ADRIAN (CH): Enhancer sequence of the 5-aminolevulinic acid synthase gene: Nucleic acid sequences mediating chemical compound induced 5-aminolevulinate synthase gene (ALAS1) expression are disclosed. Said sequences comprise at least a DR-4 binding site. Furthermore, in vitro methods for testing chemical compounds for modulation of heme and/or P 450 cytochromes synthesis are described. (Publication date: June 2, 2005) [pdf]

Interviews & general interest:

Den Besten M, Thomas AJ, Schroeder R: Life Science Research and Drug Discovery at the Turn of the 21st Century: The Experience of SwissBioGrid. Journal of Biomedical Discovery and Collaboration, Vol 4 (2009) [link]

Our experience with setting up the SwissBioGrid initative as a bottom-up collaboration for life science grid computing was worked into this experience report. Torsten Schwede and I were interviewed for this publication.

M. Podvinec: Mit Computern gegen Denguefieber. Uni Nova Wissenschaftsmagazin der Universität Basel (104) November 2006 [pdf]

This article, written in 2006, describes the virtual screening approach we applied to find new inhibitors against the dengue virus. It’s in german, and written with an interested lay audience in mind. The fact that it was written in 2006 is quite revealing about how research (sometimes) works: Only in 2010 could we publish the actual data from the project, it took that long to validate our data in experimental laboratories.

Normally, I don’t like to write too publicly about projects that aren’t done yet. In this case, we decided to go ahead and create some publicity, as there’s a download link at the end of the article, which allowed readers to contribute their computer’s free capacity to our screening efforts.

M. Podvinec: Medikamente im Test. Uni Nova Wissenschaftsmagazin der Universität Basel (93) March 2003 [pdf]

The science magazine of our university ran a special on inventions. I was contacted by their lead editor on our invention of NUBIScan, which had been published in 2002, and which our technology transfer office was  marketing at the time. After a phone discussion, I agreed to write an article about the significance of this software.

No Comments »