Michael Lam

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2023-94

May 11, 2023

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-94.pdf

It is difficult to model systems on the cellular level because of the large numbers of variables involved in these environments. The tumor microbiome is no exception. Linear models are too simplistic and unstructured to model long-term time dependencies while neural network models are so complex that it becomes difficult to gain any biological insight from the parameters. We propose using the generalized Lotka-Volterra equations, a predator-prey model that imposes an ecological structure onto the problem of inferring population-to-population interactions within the tumor microbiome. Inference was performed via model optimization on experimental data obtained from three-dimensional bioprinted tumor spheroids. We found that this deterministic method can often produce results comparable to those produced by stochastic algorithms in the past. This optimization procedure has the added benefit of being robust to experimental noise.

Advisors: Claire Tomlin

---

BibTeX citation:

@mastersthesis{Lam:EECS-2023-94,
    Author= {Lam, Michael},
    Title= {A Predator-Prey Perspective on the Tumor Microbiome},
    School= {EECS Department, University of California, Berkeley},
    Year= {2023},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-94.html},
    Number= {UCB/EECS-2023-94},
    Abstract= {It is difficult to model systems on the cellular level because of the large numbers of variables involved in these environments. The tumor microbiome is no exception. Linear models are too simplistic and unstructured to model long-term time dependencies while neural network models are so complex that it becomes difficult to gain any biological insight from the parameters. We propose using the generalized Lotka-Volterra equations, a predator-prey model that imposes an ecological structure onto the problem of inferring population-to-population interactions within the tumor microbiome. Inference was performed via model optimization on experimental data obtained from three-dimensional bioprinted tumor spheroids. We found that this deterministic method can often produce results comparable to those produced by stochastic algorithms in the past. This optimization procedure has the added benefit of being robust to experimental noise.},
}

EndNote citation:

%0 Thesis
%A Lam, Michael 
%T A Predator-Prey Perspective on the Tumor Microbiome
%I EECS Department, University of California, Berkeley
%D 2023
%8 May 11
%@ UCB/EECS-2023-94
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-94.html
%F Lam:EECS-2023-94