Imperial College London


Faculty of Natural SciencesDepartment of Mathematics

Visiting Professor



+44 (0)20 7594 2647c.colijn Website




626Huxley BuildingSouth Kensington Campus






BibTex format

author = {Plazzotta, G and Kwan, C and Boyd, M and Colijn, C},
doi = {10.1038/srep21159},
journal = {Scientific Reports},
title = {Effects of memory on the shapes of simple outbreak trees},
url = {},
volume = {6},
year = {2016}

RIS format (EndNote, RefMan)

AB - Genomic tools, including phylogenetic trees derived from sequence data, are increasingly used to understand outbreaks of infectious diseases. One challenge is to link phylogenetic trees to patterns of transmission. Particularly in bacteria that cause chronic infections, this inference is a ected by variable infectious periods and infectivity over time. It is known that non-exponential infectious periods can have substantial e ects on pathogens' transmission dynamics. Here we ask how this non-Markovian nature of an outbreak process a ects the branching trees describing that process, with particular focus on tree shapes. We simulate Crump-Mode-Jagers branching processes and compare di erent patterns of infectivity over time. We nd that memory (non-Markovian-ness) in the process can have a pronounced e ect on the shapes of the outbreak's branching pattern. However, memory also has a pronounced e ect on the sizes of the trees, even when the duration of the simulation is xed. When the sizes of the trees are constrained to a constant value, memory in our processes has little direct e ect on tree shapes, but can bias inference of the birth rate from trees. We compare simulated branching trees to phylogenetic trees from an outbreak of tuberculosis in Canada, and discuss the relevance of memory to this dataset.
AU - Plazzotta,G
AU - Kwan,C
AU - Boyd,M
AU - Colijn,C
DO - 10.1038/srep21159
PY - 2016///
SN - 2045-2322
TI - Effects of memory on the shapes of simple outbreak trees
T2 - Scientific Reports
UR -
UR -
VL - 6
ER -