You are here: Home / Loewe Lab / Members / Current Members / Loewe, Laurence

Laurence Loewe

Principal Investigator in Evolutionary Systems Biology

In Madison

Research   I   Evolvix   I   Publications   I   Posters   I   Teaching 


How does evolution change our world? How does life change in its many forms? We could solve many practical problems if we knew more. Luckily, biologists of today can combine cheap computational power with rigorous results from evolutionary genetics and current systems biology. I have decided to plunge into these opportunities, and so my research focuses on building a bridge between evolutionary genetics and molecular systems biology by building detailed computer simulation models. After >15 years in computational biology, I got so frustrated by the lack of a really good programming language for biological modeling that I've decided to develop one that makes accurate modeling easy for biologists (and many more). I am convinced that exploring this post-disciplinary rabbit hole of math, stats, computer science, and other disciplines will soon speed up my (and your?) biological research by orders of magnitude. When I'm not chasing quantitative rabbits in the various systems I'm modeling, I like sitting by the lake, contemplating the semantics of nothing, and watching movies.

 

Short CV

 

Research Interests

Evolutionary processes are at the heart of many problems that we face in our world today, ranging from antibiotics resistance evolution to species extinction. Addressing such problems requires models of the underlying causes. With my lab, I aim to improve the quality of these models by quantifying evolution with increasing precision. A key goal in this effort is to define precise maps between the biology I'm interested in and the math and computing details that I need to make the models work. I am mostly interested in mechanistic models that can help understand causal relationships in molecular and evolutionary systems. Over my career, I want to be able to combine the molecular models that I build with my population genetics models, at best a difficult challenge that becomes near impossible without long-term reproducible models that are free from semantic rot.

 

For more details see Freedman et al. (2015) The Economics of Reproducibility in Preclinical Research, PLoS Biology 13(6): e1002165, Figure 2, DOI: 10.1371/journal.pbio.1002165Thus I have taken on the challenge of creating a computing language for biology (Evolvix) that is long-term backwards compatible and hence makes computational results long-term reproducible. Long-term stability in useful computational systems is difficult (as any computer scientists can attest), and I could give a very long list of reasons why. However, I got tired of wasting my time by not having backwards compatibility because it stands in the way of my biology. So I decided to engage in some high-risk, high-reward research and create one that finally does what I expect as a biologist. In the last few years, I have developed a solid foundation and a process that can make this work. I am currently in the process of implementing this system. To spare you the hassle of trying to work with an unfinished system, I have decided to properly publish this new language only once it has reached a level of stability that I feel happy to recommend. After all, I would like to help reduce the ~$7Bn/yr of irreproducible pre-clinical data analysis in the US instead of increasing it by yet another incompatibility inducing tool....

 

At the molecular biology end, my work currently focuses on building the best possible molecular systems biology model of the Drosophila melanogaster circadian clock, incorporating all the latest data and findings. As you might guess, I want to report the resulting data in a format that is long-term backwards compatible and still reasonably easy to use. These efforts are led by Kate Scheuer and involve various other students in my lab, including undergraduates. If you are interested, please let us know. We are often looking for good students to support this.

 

Fitness Landscape

At the population genetics end, I am interested in quantifying how harmful and beneficial mutations shape the DNA sequence diversity that we can observe in populations today. This requires simulating evolving populations using realistic parameters for the strength of selection, which we estimate by population genetics methods using DNA sequence data. Unfortunately, these efforts are limited by the precision with which we can measure DNA sequence diversity, and we are currently working to see how much precision we can get for a given data set. 

 

At the EvoSysBio end, I have been working to define EvoSysBio and investigate how it could serve as a bridge between rigorous molecular systems biology models and rigorous evolutionary genetics models. I have been developing a new approach that builds on existing quantitative models from current systems biology and links them to potential incomplete fitness traits to help estimate distributions of mutational effects and fitness landscapes in silico. Since 2009, I have been organizing one EvoSysBio event each year to continually advance this long-term research program.

 

Laurence Loewe Web Portal

Other websites associated with me:

  • EvoSysBio-course.discovery.wisc.edu - Website for the 3-credit EvoSysBio Course at UW-Madison that I have been teaching each Fall since 2013. Registration is open for Fall 2016.
  • Evolvix.org - Homepage of the Evolvix model description and programming language for biology.
  • EvolutionarySystemsBiology.org - Pointer to some EvoSysBio activities, labs, and resources.
  • Evolution-at-home.org - The first global computing project for evolutionary biology. I started this in 2001, and I am currently redesigning it using Evolvix.
  • QuantBio.wisc.edu - A brief overview of some activities in quantitative biology at the University of Wisconsin-Madison.


Other webpages associated with me: