Poisson Hurdle Model-based Clustering for Microbiome Data

Join Dr. Peng Liu of Iowa State University as she proposes a model-based algorithm based on Poisson hurdle models for sparse microbiome count data. Simulation results demonstrate that the proposed methods provide better clustering results than alternative methods under various settings.

In this seminar, Dr. Liu will:

• describe an expectation-maximization algorithm and a modified version using simulated annealing to conduct cluster analysis.
• provide methods for initialization and choosing the number of clusters.
• illustrate how to apply this method using her team’s developed R package, PHclust, and apply the proposed method to a microbiome data set that results in interesting biological findings.

This method can also be applied to single-cell RNA-sequencing data.

To view upcoming speakers or recordings of past presentations, visit the CBIIT Data Science Seminar Series webpage.

The Data Science Seminar Series presents talks from innovators in the cancer research and informatics communities both within and outside of NCI.