Dear colleagues,

This is the second announcement of the Scientific Meeting this Friday, with speakers Pablo Cesar (SEN5) and Alexander Schoenhuth (MAC4). It is a lunchtime meeting, so sandwiches will be provided before the talks. We hope to see you there!

Best regards,

Willem Hundsdorfer and Ronald de Wolf

PROGRAM

Date: Friday, March 30
Time: 13.00-14.00
Location: Euler Room (Z009)
Chairman: Sander Bohte

Lecture 1, 13.00-13.30: Pablo Cesar (SEN5)
Enabling High-Quality Audiovisual Communications

Abstract: Multimedia social communication is filtering into everyday use. Videoconferencing is appearing in the living room (e.g., SkypeTV, Umi), television is becoming smart and social (e.g., Boxee, GoogleTV), and media sharing applications are transforming the way we converse and recall events (Facebook, YouTube). The confluence of computer-mediated interaction, social networking, and multimedia are reshaping social communications, bringing new challenges and opportunities. In this talk we focus on one particular problem space: domestic video conferencing. Based on a number of performance and functional requirements, we discuss about a software component that allows for dynamically manipulating and compositing audiovisual streams. This component supports high-definition video communication with low end-to-end delay. More interestingly, it is capable of dynamically reacting to low-level audiovisual cues, conversational patterns, and social interactions. We conclude the talk by introducing some enablers that are helping us in our research: the experimentation lab in the 3rd floor (Pampus room) and a number of EU-funded projects (TA2, V-conect, and Reverie).

Lecture 2, 13.30-14.00: Alexander Schoenhuth (MAC4)
Discovering individual human genetic variation with CLEVER and SMART

Abstract: Most recent advances in genome sequencing ('next-generation sequencing') have facilitated to reveal that the scale of individual variation significantly extends beyond the previously widely established assumption of only single nucleotide polymorphisms (SNPs). An estimated 8% of the human population differs in up to 500,000 nucleotides long pieces of DNA from the reference genome. Also the extent of variation in an individual which is novel relative to its ancestors is much larger than previously expected.

Despite the recent advances in sequencing technology, discovery and typing of such variation has not yet become standard. It is likely that much variation has remained undiscovered. I will present CLEVER and SMART as a combination of tools for predicting variants from next-generation sequencing data. CLEVER ('Clique-Enumerating Variant findER') organizes all sequenced pieces of DNA (reads) into a 'read alignment graph' and computes all max-cliques which reflect location-specific effects in the genome under consideration. Much of a human genome consists of repetitive sequence ('junk DNA'), which leads to disturbing ambiguities when locating the sequenced pieces of DNA. SMART ('Sparse Mixture-based Ambiguity Resolving Tool') takes care of this issue by resolving the respective ambiguities through a mixture-model based approach. Within CWI: joint work with Tobias Marschall, Gunnar Klau.