Research

The Tavaré lab has focused on stochastic computation, approximate Bayesian computation, computational biology, and statistical bioinformatics; computational cancer genomics, including evolutionary approaches to tumor heterogeneity and the study of tumors in three dimensions;  population and evolutionary genetics, including coalescent theory, statistical inference in molecular biology, human genetics, molecular evolution, and paleontology; probabilistic combinatorics. A brief summary of current projects appears below.

Current Research Projects

  • Collaborator: Dr. Daniele Biasci (Cambridge University)
  • Exploiting existing bulk and single-cell RNAseq data to identify biomarkers of response to ICI.
  • Collaborators: Prof. Rebecca Fitzgerald, Dr. Alvin Ng (Cambridge University)
  • Long-standing collaboration on the genomic aspects of EA.

Postdocs: Dr. Poly da Silva, Dr. Arash Jamshidpey (IICD)

(a) The Feller Coupling for Derangements 

  • Development of an approximation that is close in TV distance to the distribution of the cycle counts in a theta-biased derangement.
  • Among other things, very useful for simulating large derangements.

(b) Combinatorial Analysis of Some Playground Games

  • Motivated by a well-known playground game played by children in Pennsylvania. 
  • Has an interesting connection to random derangements, but also a richer combinatorial structure.
  • Postdocs: Dr. Poly da Silva, Dr. Arash Jamshidpey (IICD)
  • Collaborator: Prof. Peter McCullagh (Chicago)
  • A deeper look into sequential sampling in the ESF, with a particular focus on the behavior of the sample variance of the number of alleles found in each sample. Motivated by Fisher’s 1943 paper on species abundance.

(a) Dr. Mingzhang Yin (IICD/CS): 

  • Estimation of causal effects and relationships from observational and interventional data from cancer studies. 
  • Support: IICD/DSI postdoc, joint with Prof David Blei

(b) Russell Kunes (Statistics)

  • Statistical and ML methods for integrative analysis of scRNA and related experiments.
  • Support: NSF Graduate Research Fellowship

(c) Dr. Ignacio Vázquez-García (MSKCC/IICD), Dr. Khanh Dinh (IICD). 

  • Simulation methodology to study tumor heterogeneity, clonal development in the presence of selection, SNVs and copy number aberrations.
  • Main application is to understand the statistical behavior of DLP+ scDNA sequencing experiments.           

(d) Prof. Andy Lynch (University of St Andrews) 

  • Statistical inference for duplication rates in sequence analysis. 
  • The mutational history of a glioblastoma patient.

(e) Prof. Chin Hur, Dr. Nick Tatonetti, Dr. Jiheum Park (CUIMC)

  • Domain-knowledge informed deep learning for early detection of pancreatic cancer. 
  • Support: NIH R21

Project website: https://www.cruk.cam.ac.uk/research-groups/imaxt-laboratory

Local collaborators:

  • Long-standing research into stochastic models and statistical inference in population genetics, including statistical methods such as ABC using summary statistics such as the site frequency spectrum.
  • Applications of birth-death and coalescent processes to problems in cancer evolution.