My effective population size too large in BSP

[hojoun Jeong]

hello.

The effective population size of my bsp is too large.

I am using 80 sequences with around 10 mutations in 2300bp.

Among them, 20 sequences have exactly the same sequence.

Should I put the rate value in the gtr model?

[Alexei Drummond]

What overall rate did you use, and what are its units? If you want to read the effective population size as an effective number of individuals then you need to specify the rate in substitutions *per site* *per generation*.

If you specify the rate as substitutions per site per year (say) then the effective population size will be scaled by the generation time (in years).

If you specific the rate as substitutions per genome per generation, then the effective population size will be scaled by the number of sites in the genome, etc.

Cheers

Alexei

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beast-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/d33ac75b-bd8e-4940-a18d-5e835417083dn%40googlegroups.com.
<efe.JPG>

[hojoun Jeong]

Thank you for your reply.

I use these parameters.

Substitution rate : 1.34

clock rate : 6.7e-9

2022년 11월 14일 월요일 오후 1시 31분 13초 UTC+9에 alexei....@gmail.com님이 작성:

[Alexei Drummond]

What are the units of 6.7e-9?

To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/70a4b63e-31ed-4d64-868f-d8ca5572b390n%40googlegroups.com.

[hojoun Jeong]

its substitution per site per year.

2022년 11월 15일 화요일 오전 6시 59분 34초 UTC+9에 alexei....@gmail.com님이 작성:

[Alexei Drummond]

Then your population size estimates will be scaled to years, not number of individuals. If the generation time is 1 year they would coincide. Otherwise to get the population size in effective number of individuals you would need to divide it by the generation time in years.

Cheers

Alexei

To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/9ffb0313-fcd0-4984-bc54-8e038c031fa0n%40googlegroups.com.

[hojoun Jeong]

If the mutation rate is 1.34%/mya, what value should I put in the clock rate?

thank you.

2022년 11월 16일 수요일 오전 7시 48분 14초 UTC+9에 alexei....@gmail.com님이 작성:

[Alexei Drummond]

You should only have one rate, not two. If the rate is 1.34% per lineage per site per million years then the clock rate should be set to 0.0134 if you want branches in millions of years, or 1.34e-8 if you want branches in years. This assumes that the 1.34% is a substitutions per site per *lineage* rate and not a “divergence rate”. A divergence rate is usually defined to express the per-site difference between a *pair of taxa* separated by a common ancestor 1 million years ago (so two lineages, not one). So the substitution rate per million years would be half of divergence rate specified in that way.

But where is the 6.7e-9 coming from? That seems to be 1.34e-8/2? So maybe 6.7e-9 is actually the substitution rate, whereas 1.34e-8 is the divergence rate. In which case you should only use 6.7e-9. Then the branches will be in years. Or use 6.7e-3 if you want the branches in millions of years.

Cheers

Alexei

To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/6647e9d8-8f13-4bc5-b4e6-ad2db48d5b28n%40googlegroups.com.

[Omar Idris]

0.034 in my or .034e-6 in years? not confident myself as it gave me branch lengths in the order of e6 or 7 I think instead of multiplying it it is dividing whatever value U believe distance or substitution?

We need help.

To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/6647e9d8-8f13-4bc5-b4e6-ad2db48d5b28n%40googlegroups.com.

--

ONI