PostDoc Position(now filled!!)

Imprecise-Probability Methods for

Data Mining in Genomics and Medicine

IDSIA: Dalle Molle Institute for Artificial Intelligence,

Galleria 2, CH-6928 Manno (Lugano), Switzerland

IOSI: Oncology Institute of Southern Switzerland,

Dep. of Experimental Oncology,

Via Vincenzo Vela 6, CH-6500 Bellinzona, Switzerland


Quest. IDSIA and IOSI are inviting applications for a joint PostDoc position. We are looking for a truly outstanding person who is able to do both theoretical research and applications, and is

really interested in

.                 Imprecise probability or Bayesian robustness,

.                 Data mining, in particular classification and clustering.

 

Background information. Microarrays allow one to simultaneously measure the expression of tens of thousands of genes in human (cancer) cells, and recently also the copy-number at a hundred-thousand different positions on the genomic DNA. Typically one has to compare microarray data from as few as three, up to hundreds of patients. The data, for instance, contains information on cancer relevant genes and types of cancer. But the enormous amount of data has to be appropriately processed and integrated in order to allow the extraction of the relevant information. Novel statistical methods have to be developed that are able to cope with such complex data, yet small sample sizes. Generalizations of Bayesian methods to imprecision appear to have very good potential in this area. Imprecise-probability based classification has been already explored with good results just in the case of small sample size (and of missing data), and yet there are still plenty of research opportunities. Very little work has been carried out in the case of clustering and feature selection, so that the field is entirely open.


Job description. The goal of this project is first of all to develop new methods for clustering, classification, and feature selection with imprecise probability. Therefore the project will be mostly based initially on methodological developments at the theoretical level. The second part of the project will focus on the application of those methods to IOSIÕs genomic DNA profile datasets, primarily aiming at predicting how new patients will react to anti-cancer treatments. This prediction task will not only exploit microarray data but it will also try to integrate them with pathological and clinical data for more reliable predictions. Yet, pathological and clinical data are typically partially missing (or incomplete) and this will pose further challenges that the imprecise probability methods will have to be able to deal with.


Your background. The candidate should

.                 have excellent mathematical skills, preferable in the area of probability and statistics,

.                 have consolidated experience in data mining and with data mining/statistical packages,

.                 have good programming skills in C or C++ or Java (also R, Matlab or Mathematica could be fine),

.                 have both a Master and a Ph.D. degree in mathematics or statistics or physics or engineering or related area,

.                 be able and willing to do both theory and real applications,

.                 be interested in imprecise probability or robustness,

.                 be able of highly independent and creative research as well as able to work in group.

 

Not required but highly desirable would be

.                 experience with Bayesian networks,

.                 experience in Bayesian statistics,

.                 a background in molecular biology, genomics, or microarrays.

 

Appointment. The initial appointment will be for 2 years starting as soon as possible, with possibility of prolongation. The candidate will work in close collaboration with

.                 Marco Zaffalon (IDSIA), expert in imprecise probability and its applications to data mining,

.                 Ivo Kwee (IOSI/IDSIA), expert in applied Bayesian statistics and microarray data analysis,

.                 Francesco Bertoni (IOSI), expert in lymphomas and cancer genomics.

English is the official language at IDSIA and IOSI. The salary is around CHF 70000 per year depending on the candidateÕs profile, low taxes. There is travel funding in case of papers accepted at good conferences.

 

References. To get an idea of the research area and the required skills, please have a look at the page of the naive credal classifier 2, which is central to our imprecise-probability based research in data mining, as well at the following publications:


.                 Forconi, F., Poretti, G., Kwee, I., Sozzi, E., Rossi, D., Rancoita, P.M.V., Capello, D., Rinaldi, A., Zucca, E., Raspadori, D., Spina, V., Lauria, F., Gaidano, G., Bertoni, F. (in press). High density genome-wide DNA profiling reveals a remarkably stable profile in hairy cell leukaemia. British Journal Haematology.

.                 Agnelli L., Bicciato, S., Fabris, S., Baldini, L., Morabito, F., Intini, D., Verdelli, D., Callegaro, A., Bertoni, F., Lambertenghi-Deliliers, G., Lombardi, L., Neri, A. (2007). Integrative genomic analysis reveals distinct transcriptional and genetic features associated with chromosome 13 deletion in multiple myeloma. Haematologica 92, 56–65.

.                 Antonucci, A., Zaffalon, M. (2007). Fast algorithms for robust classification with Bayesian nets. International Journal of Approximate Reasoning 44(3), 200–223.

.                 Lombardi, K., Poretti, G., Mattioli, M., Fabris, S., Agnelli, L., Bicciato, S., Kwee, I., Rinaldi, A., Ronchetti, D., Verdelli, D., Lambertenghi-Deliliers, G., Bertoni, F., Neri, A. (2007). Molecular characterization of human multiple myeloma cell lines by integrative genomics: insights into the biology of the disease. Genes, Chromosomes and Cancer 46(3), 226–238.

.                 Corani, G., Edgar, C., Marshall, I., Wesnes, K., Zaffalon, M. (2006). Classification of dementia types from cognitive profiles data. In ECML 2006: Proceedings of the seventeenth European Conference on Machine Learning, Lecture Notes in Computer Science 4213, Springer, Berlin, 470–477.

.                 Rinaldi, A., Kwee, I., Poretti, G., Mensah, A., Pruneri, G., Capello, D., Rossi, D., Zucca, E., Ponzoni, M., Catapano, C., Tibiletti, M.G., Paulli, M., Gaidano, G., Bertoni, F. (2006). Comparative genome wide profiling of post-transplant lymphoproliferative disorders and diffuse large B-cell lymphomas. British Journal Haematology 134, 27–36.

.                 Rinaldi, A., Kwee, I., Taborelli, M., Largo, C., Uccella, S., Martin, V., Poretti, G., Gaidano, G., Calabrese, G., Martinelli, G., Baldini, L., Pruneri, G., Capella, C., Zucca, E., Cotter, F.E., Cigudosa, J.C., Catapano, C.V., Tibiletti, M.G., Bertoni, F. (2006). Genomic and expression profiling identifies the B cell associated tyrosine kinase Syk as a possible therapeutic target in mantle cell lymphoma. British Journal Haematology 132, 303–316.

.                 De Cooman, G., Zaffalon, M. (2004). Updating beliefs with incomplete observations. Artificial Intelligence 159(1–2), 75–125.

.                 Zaffalon, M., Wesnes, K., Petrini, O. (2003). Reliable diagnoses of dementia by the naive credal classifier inferred from incomplete cognitive data. Artificial Intelligence in Medicine 29(1–2), 61–79.

 

Application. Applicants should submit:


  (i) Detailed curriculum vitae, grades included,

 (ii) List of three references (including their email addresses),

(iii) Statement on how their research interests fit the above topics (1-2 pages),

(iv) Links to their theses and publications.

 

We would like to fill the position as soon as possible. Please send all documents to:

 

Marco Zaffalon, IDSIA, Galleria 2, 6928 Manno (Lugano), Switzerland.

 

Applications can also be submitted by email to zaffalon@idsia.ch (1MB max!). WWW pointers to ps/pdf/doc/html files are welcome. Use Firstname.Lastname.DocDescription.DocType for filename convention.

Please also indicate how you became aware of this announcement.

 

Deadline. The position will be filled as an adequate candidate will become available. However, we suggest to apply by 29 February 2008.

 

Thanks for your interest,

 

Marco Zaffalon, Group leader, Senior Researcher, IDSIA

Ivo Kwee, Senior researcher, Functional Genomics and Lymphoma Group, IDSIA/IOSI

Francesco Bertoni, Group leader, Functional Genomics and Lymphoma Group, vice-director, Laboratory of Experimental Oncology, IOSI.

[Ticino, January 2008]

 

 

 

ABOUT IDSIA. The ÒDalle MolleÓ Institute for Artificial Intelligence is a joint research institute of University of Lugano and Swiss Italian University for Applied Science.

Our research focuses on uncertain reasoning, imprecise probabilities, graphical models, data mining, artificial neural nets, reinforcement learning, complexity and generalization issues, unsupervised learning and information theory, forecasting, artificial ants, combinatorial optimization, evolutionary computation.

IDSIAÕs group on imprecise probability is currently made of a group leader, two postdocs and two Ph.D. students.

IDSIA is small but visible, competitive, and influential. The "X-Lab Survey" by Business Week Magazine ranked IDSIA among the world's top ten labs in Artificial Intelligence. IDSIA's algorithms hold the world records for several important operations research benchmarks (see Nature 406(6791):39-42 for an overview of artificial ant algorithms developed at IDSIA).

IDSIA develops and participates in projects funded by the Swiss National Science Foundation, by the European Commission and by the Swiss Commission for Technology and Innovation. IDSIA also consults private firms and enterprises.

IDSIA (see our building) is located near the Swiss supercomputing center and close to the beautiful city of Lugano (see below).

 

ABOUT IOSI. The Laboratory of Experimental Oncology of the Oncology Institute of Southern Switzerland (IOSI) employs a group of scientists working in the area of molecular pharmacology, drug development, cancer genetics and molecular biology. The laboratory is particularly active in the study of the biology and pathogenesis of lymphomas and prostate cancer. These studies promise to shed light into basic mechanisms of tumorigenesis and provide insights for development of novel therapeutic and preventive strategies for cancer. The  lymphoma and functional genomics unit currently consists of investigators with expertise in the areas of molecular biology, cytogenetics and microarray analysis. The unit performs gene expression and arrayCGH studies, in collaborative projects with  the the main international cancer Centers All the equipment needed for the most advanced genetic and molecular biology-based research is available, including the Affymetrix GeneChip system. The Oncology Institute of Southern Oncology, directed by Prof. Franco Cavalli, is the world leading group for lymphomas and clinical trials for new anti-cancer drugs. The IOSI Laboratory of Experimental Oncology is located within the Institute for Research in Biomedicine in Bellinzona.

 

LUGANO and BELLINZONA are located in Ticino, the scenic southernmost province of Switzerland, origin of special relativity and the WWW. Milan, Italy's center of fashion and finance, is 1 hour away, Venice 3 hours. Switzerland boasts the highest citation impact factor, the highest supercomputing capacity pc (per capita), and the most Nobel prizes pc (450% of the US value).