Speaker: Prof. Dr. Sergei L. Kosakovsky Pond, Temple University, Dep. of Biology, Philadelphia, USA

Title: Beyond software tuning: scaling up comparative coding sequence analysis using approximations and models that adapt their complexity to the data (Presentation in English)
Date: Monday, 28 May 2018, 11:00 a.m.
Location: Carl-Bosch-Auditorium, Studio Villa Bosch, Schloss-Wolfsbrunnenweg 33, 69118 Heidelberg (Studio entrance between Villa Bosch and HITS)
Parking: Parking garage "Unter der Boschwiese" (free of charge)

Genetic sequence data are being generated at an ever-increasing pace, while many analytical techniques that are commonly used to make biologically meaningful infer-ences on these data are still “stuck” in the “small data” age. For example, a practical upper bound on the number of sequences that can be analyzed with many popular comparative phylogenetic methods is 1000, especially if codon-substitution models are used. These types of models are an essential tool for deciphering the action of natural selection on genetic sequences, and have been used extensively in biomedical and basic science applications, for example to quantify pathogen evolution: drug re-sistance, zoonotic adaptation, immune escape.

We show how his number can be raised by several orders of magnitude, enabling in-depth study of gene-sized alignments with 10000 – 100000 sequences, much more extensive model testing, or the implementation of more realistic models with added complexity. This can be accomplished via an adaptation of machine learning tech-niques originally developed in the context of large-scale data mining (latent Dirichlet allocation models), and for variable selection.

Specifically, we describe a relatively general approximation technique to limit the num-ber of expensive likelihood function evaluations a priori, by discretizing a part of the parameter space to a fixed grid, estimating other parameters using much faster sim-pler models, and integrating over the grid using MCMC or a variational Bayes ap-proach. We demonstrate how this technique can achieve 100× or greater speedups for detecting sites subject to positive selection, while improving statistical performance. Other analyses where there are only a 2-3 parameters of interest (e.g. detection of directional selection in protein sequences) can be accommodated. When discretization is not appropriate, it is often possible to develop methods that employ variable para-metric complexity chosen with an information theoretic criterion. For example, in the Adaptive Branch Site Random Effects model, we quickly select and apply models of different complexity to different branches in the phylogeny, and deliver statistical per-formance matching or exceeding best-in-class existing approaches, while running an order of magnitude faster.

Curriculum vitae: Please see: http://spond.github.io/CV.js/cv.html

Contact: Benedicta Frech (This email address is being protected from spambots. You need JavaScript enabled to view it., phone: 06221-533-263)

Am Freitag, den 17. November 2017, um 20:00 Uhr findet der Vortrag „Die Kunst der Bewegung – Von Exo-Skeletten und humanoiden Robotern“ von Prof. Dr. Katja Mombaur im Rahmen des „International Science Festival – Geist Heidelberg“ im DAI Heidelbrg statt. Frau Prof. Mombaur untersucht mit ihrer Arbeitsgruppe an der Universität Heidelberg menschliche Bewegungen im Alltag und beim Sport. Elementar dafür sind Bewegungsmessungen am Menschen selbst, mathematische Modelle und Simulationen am Computer. Eines von Frau Mombaur Zielen ist, Exo-Skelette zu entwickeln, die beim Laufen helfen oder die Wirbelsäule entlasten. Außerdem leitet Frau Mombaur das Heidelberg Center for Motion Research, an dem sich Natur- und Geisteswissenschaftler gemeinsam u. a. mit der Verbindung von Bewegung, Psyche, Kognition und Körper beschäftigen.

Der Eintritt beträgt im Vorverkauf 8 Euro, ermäßigt 5 Euro.

Mehr Informationen erhalten Sie auch unter https://dai-heidelberg.de/de/veranstaltungen/katja-mombaur-joerg-troeger-17049/.

Speaker: Prof. Dr. Peter V. Coveney, Computational Chemistry Section, University College London, UK

Title: The Virtual Human: In Silico Methods for Personalised Methods (Presentation in English)

Date: Monday, 20 November 2017, 11:00 a.m.
Location: Carl-Bosch-Auditorium, Studio Villa Bosch, Schloss-Wolfsbrunnenweg 33, 69118 Heidelberg (Studio entrance between Villa Bosch and HITS)
Parking: Parking garage "Unter der Boschwiese" (free of charge)

The era of personalised medicine offers at once a huge opportunity and a major challenge to computational science. The potential impact centres around our ability to marshal substantial quantities of patient data and to use them to perform predictive, mechanistic modelling and simulation in order to deliver therapies and to enhance clinical decision making, on time scales which are far shorter than those usually considered in the context of academic research and development activities. Secure access to personal data, as well as to powerful computational resources, is essential. I shall provide a couple of examples which illustrate the current state of the art. One addresses clinical decision support in the context of blood flow within neuro-vascular pathologies; the other is concerned with patient specific drug discovery and treatment. We shall discuss the underlying e-infrastructure requirements, including data, compute and networks, and reflect on the potential for cloud and other forms of e-infrastructure provision to meet the anticipated future demand for resources.

Curriculum vitae:
Prof Peter V. Coveney holds a chair in Physical Chemistry, is an Honorary Professor in Computer Science at University College London (UCL) and is Professor Adjunct at Yale University School of Medicine (USA). He is Director of the Centre for Computational Science (CCS) at UCL. Coveney is active in a broad area of interdisciplinary research including condensed matter physics and chemistry, materials science, as well as life and medical sciences in all of which high performance computing plays a major role. He has led many large scale projects, including the EPSRC RealityGrid e-Science Pilot Project (2001-05) and its extension as a Platform Grant (2005-09); he is also PI on several current grants from EPSRC and other agencies, including the the role of Coordinator of the EU FP7 Virtual Physiological Human (VPH) Network of Excellence (2008-13). He has been the recipient of many US NSF and DoE as well as European supercomputing awards (from DEISA and PRACE), which provide access to several petascale computers. Coveney chaired the UK Collaborative Computational Projects Steering Panel (2005-15) and has served on programme committees of many conferences, including the 2002 Nobel Symposium on Self-Organisation; he was Chair of the UK e-Science All Hands Meeting 2008, and of the Discrete Simulation of Fluid
Dynamics conference 2003. He has published more than 400 scientific papers and co-authored two best-selling books (The Arrow of Time and Frontiers of Complexity, both with Roger Highfield) and is lead author of the first textbook on Computational Biomedicine (Oxford University Press, 2014). Coveney is a founding member of the UK Government’s E-Infrastructure Leadership Council and a Medical Academy Nominated Expert to the UK Prime Minister's Council for Science and Technology on Data, Algorithms and Modelling which has led to the creation of the London based Turing Institute.

Benedicta Frech (This email address is being protected from spambots. You need JavaScript enabled to view it., phone: 06221-533-263)

Für alle Studierenden, die nach Heidelberg in einen Studiengang der Fakultäten für Mathematik und Informatik sowie Physik und Astronomie wechseln, bietet die Fachschaft MathPhysInfo eine Informationsveranstaltung und eine Kneipentour an, um den Studieneinstieg in Heidelberg zu erleichtern.
Der Termin und Treffpunkt sowie weitere Informationen sind auf den Seiten der Fachschaft MathPhysInfo unter [1] zu finden.

[1]  https://mathphys.fsk.uni-heidelberg.de/w/veranstaltungen-fuer-wechsler_innen

Speaker: Prof. Dr. Gerhard Jäger, University of Tübingen, Institute of Linguistics

Title: From Words to Features to Trees: Computing a World Tree of Languages from Word Lists
Presentation in English

Date: Monday, 16 October 2017, 11:00 a.m.
Location: Carl-Bosch-Auditorium, Studio Villa Bosch, Schloss-Wolfsbrunnenweg 33, 69118 Heidelberg (Studio entrance between Villa Bosch and HITS)
Parking: Parking garage "Unter der Boschwiese" (free of charge)

Since over 200 years, historical linguists strive to reconstruct family trees of human languages using systematic comparisons of vocabulary and grammar of extant or documented languages. Since about 20 years, these efforts are complemented by computational approaches, deploying phylogenetic inference algorithms from computational biology to analyse language data. So far, both lines of research have been confined to individual language families, i.e., phylogenetic units with a time depth of at most 10,000 years.
In this talk I will present and discuss a workflow that starts out from unannotated word lists from ca. 6,000 languages and dialects across the world. Using feature extraction techniques from machine learning, a feature matrix is extracted which in turn serves as input for Maximum-Likelihood phylogenetic inference (using the software RAxML). This leads to a phylogenetic tree over those languages and dialects, which is in very good agreement with expert classifications, correlates well with anthropological and genetic data, and also reveals some interesting deeper signals.

Curriculum vitae:
Gerhard Jäger (http://www.sfs.uni-tuebingen.de/~gjaeger/) is professor of General Linguistics at Tübingen University. He received his PhD and habilitation from Humboldt University at Berlin and held previous positions at Munich, UPenn, Utrecht and Stanford. He is PI of an ERC Advanced Grant "Language Evolution: The Empirical Turn" and co-PI of the interdisciplinary DFG-Research Unit "Words, Bones, Genes, Tools: Tracking Linguistic, Cultural and Biological Trajectories of the Human Past". His research interests include computational historical linguistics and game-theoretic pragmatics.

Benedicta Frech (This email address is being protected from spambots. You need JavaScript enabled to view it., phone: 06221-533-263)

The Institute of Computer Engineering (ZITI) at Heidelberg University, Germany, invites applications for a newly created

W3-Professorship in Medical Technology.

The core competencies of the ZITI lie in the development and operation of specialized computer hardware and hardware-associated soft- and firmware for the solution of challenging computer problems and for data acquisition, in the development of application specific microchips (analogue and digital) for digital communication and sensors, as well as in the areas of automation, robotics and biomechanics.

The main focus of research of this new professorship should lie in the development of new concepts for and in the construction of medical instruments, to which the core competencies of ZITI can contribute. For their practical application and development, the selected candidate should establish cooperation with both clinical faculties of the University and the German Cancer Research Center (DKFZ), where already many developments and applications of new technologies have occurred (e.g. PET/MRT, 7T, radiotherapy, proton and ion therapy, robotics for surgery, prosthesis).

The professorship will be in the Department of Physics and Astronomy.

The successful candidate will be expected to contribute to teaching within the master programme „Computer Engineering“ as well as to the Bachelor and Master programmes in physics, in particular for the specialization “Medical Physics”.

Prerequisites for application are a university degree and (in accordance with Article 47, paragraph 2 of the Higher Education Law of the State of Baden-Württemberg) a Habilitation, a successfully evaluated junior professorship or equivalent qualification. The University of Heidelberg seeks to achieve a higher proportion of women in areas where they have not been adequately represented so far. Women with the required qualification are particularly encouraged to apply. Preference will be given to disabled applicants with equal qualifications for the position.

Qualified candidates are invited to submit their application including CV, a list of publications, a list with teaching experience and a research plan in electronic form (pdf) until 16.10.2017 to the Dean of the Department of Physics and Astronomy, Im Neuenheimer Feld 226, D-69120 Heidelberg (This email address is being protected from spambots. You need JavaScript enabled to view it.). Further information can also be obtained from the managing director of ZITI (This email address is being protected from spambots. You need JavaScript enabled to view it.).


back to top