Uncategorized
The STATegra Summer School in Omics and Data Integration has been held in Benicassim, Spain, from the 7th to the 11th of September 2015, providing training to more than 30 among experimentalists and bioinformaticians in the analysis of multi-omics experiments.
Further information on the summer school website: http://stategrasummerschool.com/
Uncategorized
The second edition of the “Statistical Methods for Omics Data Integration and Analysis” conference has been held in Valencia, September 14-16, 2015.
MXM has participated with a talk regarding application of data-integration methods for the analysis of public microarray data in Leukemia patients.
The program of the conference and further information on SMODIA2015
website: http://smodia2015.cipf.es/
Uncategorized
Speaker: Karen Sachs, Stanford University School of Medicine
Date: 03 September 2015 Time: 11:00 – 13:00
Location: Seminar Room I, FORTH, Heraklion, Crete
Host: Ioannis Tsamardinos
We have previously introduced the application of probabilistic models for elucidation of statistical relationships from single cell proteomic data (Sachs et al, Science 2005), an approach enabled by the key insight that each cell may be considered an observation of the underlying biological system. The application of this approach has been partially limited by the low dimensionality of the data modality (flow cytometry), which enabled measurements of only up to ~10 proteins of interest per cell. In 2011, we introduced a next generation single cell proteomic technology, mass cytometry or CyTOF (cytometry time of flight, Bendall et all, Science 2011), which enables quantification of 30-40 parameters per cell.
In this talk, I will describe the CyTOF technology and illuminate its advantages and potential pitfalls. I will then discuss my perspective on causal modeling of single cell data and what hurdles remain in this application.
Her work involves probabilistic modeling of signal transduction pathways. She uses probabilistic approaches to study how different proteins in a signaling pathway depend upon and influence each other.
This information is extracted from data, in which these proteins have been measured many times, each time in a single cell, using a technology called “flow cytometry”.
Uncategorized
From July 1st to July 3rd, Yannis Pantazis attended the workshop “Mathematical Trends in Reaction Network Theory” at Copenhagen, Denmark. He presented a poster entitled “Pathwise Information-theoretic Metrics for Parametric Sensitivity Analysis” where the relative entropy rate and the associated path Fisher information matrix were demonstrated as a means of assessing the parameters’ sensitivities.
Uncategorized
Speaker: Alessio Farcomeni, Assistant Professor in Medical Statistics at Sapienza University of Rome
Date: 30 April 2015 Time: 12:00 – 13:00
Location: Seminar Room I, FORTH, Heraklion, Crete
Host: Ioannis Tsamardinos
Issues about the global level arise whenever many hypotheses are tested at once. A common approach is to use Bonferroni correction, which has very low power when a large number of tests are performed.
Many current problems (e.g., gene discovery) involve a number of tests in the order of the tens or even hundreds of thousands. An alternative to Bonferroni is control of the False Discovery Rate, the expected proportion of false rejections over the number of discoveries, if any. The problem of extending FDR control to dependent test statistics is still somewhat open and partly confined to the case of positive dependence. Ad-hoc modifications to the standard procedure often lead to loss of power. We show that a certain degree of dependence, no matter its direction, is allowed among the test statistics without need for any modification to the standard procedures controlling the FDR.
We then introduce a new class of multiple testing procedures which can control the FDR and other functionals of the ratio of false rejections and discoveries at the user’s choice. The class is based on the idea of generalized augmentation: at first hypotheses are rejected without any correction, then the set of rejections is adjusted by adding or removing a suitable number of rejected hypotheses. We focus on the probability of FDP exceeding a pre-specified threshold; referred to as False Discovery eXceedance, FDX. The new class is seen in simulations and real data applications to be much more powerful than the available methods in controlling the FDX, especially when the number of tests is large. A slight modification makes Type I error rate control valid under arbitrary dependence of the test statistics. Different applications are used to illustrate the use of FDP control in multiple testing, including: spatial disease mapping, gene discovery, image reconstruction.
Alessio Farcomeni is a statistician, working as assistant professor in Medical Statistics at Sapienza – University of Rome. His work focuses on robust statistics, longitudinal models, categorical data analysis, cluster analysis, and multiple testing. He also is involved in clinical, ecological, and econometric research.
Uncategorized
Our work has been awarded with the Boeringer Ingelheim Cancer Research Award (Best Abstract / Poster of the Year) 2014 from the Norwegian Lung Cancer Group. This is joint work with our NTNU collaborators involving the analysis of case-control molecular profiles of lung cancer from the HUNT biobank. For more information see http://www.nlcg.no/node/122. The poster presented early research results of this on-going project, supporting the evidence that there is diagnostic information for lung cancer and mesothelioma in human serum. The full citation of the poster is:
Robin Mjelle, Trygve Andreassen, Vincenzo Lagani, Konstantinos Kerkentzes, Ioannis Tsamardinos, Tone Bathen, Erik Larsson, Oluf Dimitri Røe, ““HUNTing” early diagnostic biomarkers for lung cancer / mesothelioma : a pilot analysis from the Cancer-Biomarker-HUNT study”
Stay tuned for more upcoming results on this very promising, highly exciting project that could yield early prognostic and diagnostic markers of thoracic cancers.
Uncategorized
As of 1-1-2015 the Bioinformatics Laboratory of the Institute of Computer Science at the Foundation for Research and Technology, Hellas has been merged with the Computational Medicine Laboratory to yield the Computational BioMedicine Laboratory (http://www.ics.forth.gr/cbml ) . The new laboratory’s Head is Dr. Kostas Marias.
As the Head of the former Bioinformatics Laboratory, I (Prof. Tsamardinos) would like to thank the members of the lab, that gave their 100% to make it a highly successful (even if short lived) lab. I am forever in debt to them for their continuous efforts, grateful for their choice to join despite the risks and uncertainty that it entailed, and honored to have been the Head of the group.
The Bioinformatics Laboratory was founded in 2011 with only 16K euro in funding. During the 4 years of the lab (2011-2014), we attracted more than 1.1M euro, reached a maximum size of 19 personnel, got more than 2000 citations to our published work, published papers in forums ranging from ICML, UAI, JMLR to NAR and Cell Metabolism and presented invited talks to prominent Universities, groups and conferences around the world.
A few words about the future and how the merge affects the MXM group. The merge presents new opportunities for the MXM members to collaborate with the other members of the CBML laboratory, including joint publications and funded projects which we intend to explore. In addition, it reduces the administrative load on Prof. Tsamardinos. The new ERC grant CAUSALPATH that just started this year is a major current activity of the MXM group and is hosted in the University of Crete. Thus, a large part of the MXM research activities will be de-facto moving in the University of Crete.
We would like to extend our best wishes to the new laboratory and its Head and express our eagerness to contribute to its success.
Uncategorized
On behalf of the STATegra consortium our group participated in the organization of a workshop titled “Statistical Methods for Omics Data Integration and Analysis, SMODIA”. The workshop took place in November 10-12, 2014 in Heraklion, Crete, Greece. It was organized as part of the dissemination efforts of the STATegra FP7 EU project, titled “User-driven Development of Statistical Methods for Experimental Planning, Data Gathering, and Integrative Analysis of Next Generation Sequencing, Proteomics and Metabolomics data”. The workshop included an impressive panel of invited speakers, such as Dr. Jesper Tegner from Karolinska Institute, Dr. Andrew Teschendorff from UCL, London and Dr. Sven Nelander from Uppsala University. Workshop presentations trigger interesting ideas and questions which were discussed during the panel meeting, where all members of the STATegra consortium participated. Finally, due to workshop’s high impact and participation, SMODIA will held again next year and the venue will announced shortly.
The SMODIA 2014 Organizing Committee
Uncategorized
Eleni Sgouritsa presents her work on Tuesday, November 4th , 12:00pm, Seminar Room I, Institute of Computer Science, Foundation for Research and Technology Hellas (FORTH).
Title:
Identifying confounders and telling cause from effect using latent variable models
Abstract:
Drawing causal conclusions from observed statistical dependencies is a fundamental problem. Conditional independence-based causal discovery (e.g., PC or FCI) cannot be used in case there are no observed conditional independences. Alternative methods investigate a different set of assumptions, namely restricting the function class, e.g., additive noise models. In this talk, I will present two causal inference methods employing different kind of assumptions than the above.
The first is a method to infer the existence and identify a finite confounder of a set of observed dependent variables. It is based on a kernel method to identify finite mixtures of nonparametric product distributions. In the second part, I will focus on the problem of causal inference in the two-variable case (assuming no confounders). A known assumption is used, namely that if X causes Y, P(X) contains no information about P(Y|X). In contrast, P(Y) may contain information about P(X|Y). This asymmetry has implications for common machine learning tasks such as semi-supervised and unsupervised learning, which are employed by the proposed causal inference method.
Short Bio:
Eleni Sgouritsa is in the last year of her PhD at the Max Planck Institute for Intelligent Systems in Tübingen, Germany, under the supervision of Dominik Janzing and Bernhard Schölkopf. She is a member of the causality group which aims at developing novel methods for inferring causal relations from data. Prior to this, she received an MSc and a BSc degree from the Univ. of Houston and the Univ. of Athens, respectively.
Uncategorized
Pr. Ioannis Tsamardinos is invited to give a talk at MPI for Intelligent Systems in Tuebingen.
Title: Advances in Integrative Causal Analysis
Abstract: We’ll present the concept and approach of Integrative Causal Analysis(INCA), where causal models and relations are induced from multiple,heterogeneous datasets. The datasets maybe heterogeneous in terms of the variable sets they measure or the experimental conditions (interventions)under which they were collected. To solve the problem, all causal models consistent with all datasets are induced and their invariant characteristics (causal features) are found. We introduce a technique where the problem is converted to a SAT (satisfiability) problem that exploits decades of research in satisfiability engines for improved efficiency. In addition, it is also shown that INCA algorithms can in principle also be extended to provide quantitative predictions. The general idea is tested on a couple of dozen real datasets where it is shown that the correlation between two variables never jointly measured, can be predicted based on causal models induced by 2 datasets with overlapping variables. The problem of resolving conflicts arising during learning is also dealt with as well as the problem of identifying the parts of the models that are of high confidence. Finally, new projects and directions are discussed and particular the application of these techniques to learning biological signal pathways from mass cytometry data.