BSP for calculation of the size of viruses or people infected with viruses

[Lu Ludwig]

Dear BEAST users, I studied evolution of hepatitis viruses and have
practised BEAST for several times. Now I have a simple question. In the
Rough Guide to BEAST 1.4, BSP is available in the Tree priors and designed
to calculate the effective breeding population size through time. So I
consider that BSP can depict the change of hepatitis viruses population
based on my sampling time (e.g, 1840 to 2007). But I read the
"Investigating the origin and spread of hepatitis C virus genotype 5a" in J
Virol, 2006, in which BSP depicted the change in the effective number of
infected individuals through time. Then I was coufused. Dose BSP describe
the population of the viruses or the population of the people infected with
the viruses?

If BSP can depict the change of the population of the people infected with
the viruses, what is the principle or theory to account for it?

Does anyone could help me? Thank you very much.

I am very sorry for my poor English.

Lu Yi Han
Department of Epidemiology, School of Public Health
Fudan University
Shanghai, China

_________________________________________________________________
与联机的朋友进行交流，请使用 MSN Messenger: http://messenger.msn.com/cn

[alexei....@gmail.com]

Hi Lu,

If you have sampled a single virus sequence from a random sample of
infected individuals then theory tells us that the estimated
population size should most closely reflect the number of infected
individuals rather than the total number of viruses. Although this has
been asserted for a long time, the first paper to *prove* that a
coalescent process (virus population) inside another coalescent
process (host population) is itself a coalescent process was:

Welch, Nicholls, Rodrigo, Solomon (2005) Integrating genealogy and
epidemiology: The ancestral infection and selection graph as a model
for reconstructing host virus histories. Theoretical population
biology.

If, one the other hand you have more than one sequence from each
patient then the picture is cloudy, as your genealogy will have some
coalescent events within infected individuals and some between
infected individuals. In this case there is no simple interpretation
of the Bayesian skyline plot.

Finally if you have a number of virus sequences sampled from a single
infected individual, then the population size estimates will be of the
virus population within that individual.

Cheers
Alexei