Skip to content

NCBI-Codeathons/beyond-phylogenies-team5

Repository files navigation

Team 5: Estimating direct transmission in routine clinical sequencing from well sampled locations

List of participants and affiliations:

  • Abhishek Mishra
  • Ashwin Babu
  • Chrispin Chaguza
  • Joshua Levy
  • Karthik G
  • Lax Iyer
  • Tyler Chefin
  • Xiaopeng Bian

Project Goals

  • Benchmark three existing tools and include a simple SNP threshold as a baseline
    • Create simulated dataset and curate real world dataset.
    • Run tools on both datasets.
    • Compare estimates across tools for both datasets.
    • Identify a tool(s) based on estimates and scalability.
  • Run selected tool(s) on dataset from a single well sampled location
  • Visualize estimates

Approach

workflow

We selected three well-used tools to be part of our benchmarking analysis: TransPhylo, outbreaker2, and transcluster. We will include a simple SNP threshold to identify transmission clusters as a baseline.

We simulated an outbreak using TransPhylo and simulated sequences from the resulting phylogenetic tree using phastsim. For a real world dataset, we used all the sequences from Massachusetts, USA from the form of July 2010 from GISAID. This time period covers a well characterized superspreading event (Ref).

Results

ROC ROC subset

Future Work