Research Focus Class

Our Research Focus Classes (RFC) are a special kind of lecture: they are more interactive and research oriented than typical lectures. Students participating in an RFC should be aware that they are not only getting in touch with real research but also have to expect doing independent work.

Overview RFC SS 2016: Secure Genome Analysis

• SWS: V3/Ü2, ECTS: 6
• Study programs: Master Informatik (Software und Kommunikation), Master Software Systems Engineering (Communication), Master Media Informatics
• Dates: to be determined with the participants
• Due to limited capacity, prior registration is required!
• Registration and questions: contact Martin Henze

Motivation: Why Secure Genome Analysis?

Genome analysis is a main enabler for medical and pharmaceutical research and exciting new medical services. It is used, e.g., to identify congenital diseases, to identify persons or their ancestry, or to check a person for other characteristics such as susceptibility to certain environmental influences.

However, personal genomics and genomic research face a huge privacy challenge: On the one hand, research institutes and companies are in dire need of real DNA information to conduct their reserach. On the other hand, users have serious concerns about sharing their DNA information with untrusted third parties. These opposing interests are a severe impediment to advancing genomic research and personal service models based on genomics.

Hence, solutions that rigorously guarantee security and privacy when processing genomic information need to be developed.

What is the class about?

The goal of this course is to design, implement and submit a solution (or several) to the iDash competition. This competition poses three tasks concerning secure genome analysis. As a team, we will pick one or two tasks and solve them in the course of the RFC.

In this course you will, ...

• learn the basics of secure computations hands on, using advanced cryptographic tools and frameworks,
• learn basic approaches to secure genomic analysis,
• design and implement your own solution to one of the competition tasks,
• build quick prototypes to try different approaches 
• benchmark and optimze your solution for speed (and possibly other aspects),
• inspire your fellow participants with your ideas and be inspired by them,
• contribute towards one joint solution,
• hopefully, win this comptetition, as a team!

 

What should you bring?

This class will serve as an introduction on how to conduct research in communication systems. Besides learning about the actual topics, taking this course is an ideal preparation for doing a master thesis in our group. The course is only open for Master students. You should have prior knowledge at least in the basics of data communication and an interest in doing independent research. As this course is supposed to be highly interactive, seats are limited. Should we receive more registrations than seats available, we will select students based on their qualification for this course.

We do not require any special skills or knowledge. Certainly, helpful are the following

• Preliminary knowledge or strong interest in cryptography
• Quick (and dirty) coding skills (e.g., in Python, C++)
• A very result-oriented way of working
• The willingness to contribute time

Structure of the RFC

The RFCs are research-oriented courses following an interactive schema. To do so, we give a short introductory lecture about the topic (3-5 lecture slots). Afterwards, we dive right into the competition, where the progress and next steps are discussed every one or two weeks in a joint meeting and each participants works individually in between. The RFC ends with the submission of the final solution(s) to the competition and a short joint presentation of the achieved results.

Overall, we have the following rough time schedule, which is oriented towards the competitions own timeline (http://idashworkshop2016.weebly.com/about.html):

• 15.03.2016: Competition tasks are released on homepage
  • idashworkshop2016.weebly.com/competition-tasks.html
  • Potential tasks are chosen
• Lecture Phase: April (probably 3-4 lectures in 2-3 weeks)
  • Getting up to speed!
  • Learn about examples, approaches, tools, ...
  • Learn about last year's best solutions
• Exploration and trial phase: May - mid-July
  • Chose your task
  • Establish baselines
  • Try different approaches and directions
  • Benchmark, benchmark, benchmark
• Solution phase: mid-July - August
  
  • Combine the best approaches and ideas towards a joint solution
  • Optimize, polish and fine tune
• Submission: End of August
  • Submit joint solution to chosen tasks
• Presentations: September
  • Present your work to the COMSYS group
• Notification: Early October
  
  • Winners are announced.