Survival of the Smartest: Sparse Recovery in Biology

Series: 
Applied and Computational Mathematics Seminar
Monday, October 5, 2015 - 14:00
1 hour (actually 50 minutes)
Location: 
Skiles 005
,  
Mathematisches Institut Lehrstuhl für Mathematische Optimierung
,  
Survival can be tough: Exposing a bacterial strain to new environments will typically lead to one of two possible outcomes. First, not surprisingly, the strain simply dies; second the strain adapts in order to survive. In this talk we are concerned with the hardness of survival, i.e. what is the most efficient (smartest) way to adapt to new environments? How many new abilities does a bacterium need in order to survive? Here we restrict our focus on two specific bacteria, namely E.coli and Buchnera. In order to answer the questions raised, we first model the underlying problem as an NP-hard decision problem. Using a re-weighted l1-regularization approach, well known from image reconstruction, we then approximate ”good” solutions. A numerical comparison between these ”good” solutions and the ”exact” solutions concludes the talk.