TNet-Geo is a tool for geographical transmission network analysis when multiple strain sequences from different infected hosts are available from the different geographic regions (e.g., countries) under consideration. TNet-Geo is parameter-free and highly scalable and can be easily applied within seconds to datasets with thousands of strain sequences from hundreds of geographical regions. It is built upon the TNet tool, which uses multiple strain sequences per infected host to infer a host-to-host transmission network. We refer the reader to https://compbio.engr.uconn.edu/software/tnet/ for further details on TNet.
TNet-Geo takes as input a phylogeny of the sampled strain sequences and analyzes it to estimate the underlying geographical transmission network. In particular, it labels each internal node of the strain phylogeny with a geographic region label, and this labeled phylogeny can then be used to infer how the infectious disease was transmitted across regional boundaries. The underlying assumption on which TNet-Geo is based is that the transmission of infection is easier within a geographical region than between two different geographical regions. Note that there are some fundamental differences between traditional (host-to-host) transmission networks and geographical transmission networks. Most importantly, in a traditional transmission network, the question of interest is “Who infected whom?”, while in geographical transmission networks, the relevant question is “What was the extent of infection spread from one region to another in a specified time frame?”. Accordingly, the output from TNet-Geo has to be interpreted differently compared to outputs from traditional transmission network inference tools.
TNet-Geo was implemented by Saurav Dhar and is available open source under the GNU general public license. TNet-Geo is implemented in Python 3. It also uses the Biopython library that can be freely downloaded as described at https://biopython.org/.
- Manual: TNet-Geo_Manual.pdf
- Source code: https://github.com/sauravdhr/tnet_geo (available open source under GNU GPL)
TNet-Geo can be cited as follows:
- TNet: Transmission Network Inference Using Within-Host Strain Diversity and its Application to Geographical Tracking of COVID-19 Spread
Saurav Dhar, Chengchen Zhang, Ion Mandoiu, Mukul S. Bansal.
IEEE/ACM Transactions on Computational Biology and Bioinformatics, 19(1): 230-242, 2022. (Published online in July 2021.)
→ The global COVID-19 dataset used in this paper is available from the Datasets Page
Acknowledgement tables for COVID-19 sequences from GISAID used in our analysis: Global sequences and US sequences.
Funding: Development of the software resource(s) available from this webpage was funded in part by U.S. National Science Foundation award CCF 1618347.