Decide: DNA sequence length: 1000 or 5000?

[richelbilderbeek]

Note to self first, will discuss with Giappo later.

In this article

@article{sarver2019choice,
  title={The choice of tree prior and molecular clock does not substantially affect phylogenetic inferences of diversification rates},
  author={Sarver, Brice AJ and Pennell, Matthew W and Brown, Joseph W and Keeble, Sara and Hardwick, Kayla M and Sullivan, Jack and Harmon, Luke J},
  journal={PeerJ},
  volume={7},
  pages={e6334},
  year={2019},
  publisher={PeerJ Inc.}
}

in SupMat 2 (PDF) they find:

If we are convinced, we could use 5k as well.

[richelbilderbeek]

The reason 5k is used, is from [1], where estimated speciation rates are not getting better after 5k nucleotides. Because razzo does not estimate speciation rates, we can AFAICS use 1k nucleotides regardless (as the experiment takes up a lot of runtime) and mention it someplace.

• [1] Wertheim, Joel O., and Michael J. Sanderson. "Estimating diversification rates: how useful are divergence times?." Evolution: International Journal of Organic Evolution 65.2 (2011): 309-320.

[richelbilderbeek]

Sent email to @Giappo:

razzo approve/suggest: DNA sequence length

6 January 2020 14:55 3 KB From: Richel Bilderbeek To: Giovanni Laudanno

[...]

I need your approval/suggestion of the DNA sequence length we use in razzo.

Currently, in razzo, we use a DNA sequence length of 1k. A similar research [2] uses a DNA sequence length of 5k, because in [3] they found that using a DNA alignment longer than 5k, the estimation of the speciation rate does not improve.

I'd say, because we do not estimate a speciation rate, we can ignore (yet mention) the practice of using 5k nucleotides, to gain more replicates instead.

So, do you approve?

• 1. Yes, I approve
• 2. No, I need more information/options

[...]

[richelbilderbeek]

Will be decided on meeting Friday/tomorrow.

[Giappo]

Yes I approve.