Is there gold standard data for ancestral reconstruction?

Question

[c_u]

Is there gold standard data for ancestral reconstruction? I am asking because I am trying to create a new approach for ancestral (gene/protein sequence) reconstruction (using a algorithm that has not been used before for this) and I wanted to compare the performance of current tools like MrBayes, MEGA, Mesquite etc.

Is there a way to compare their performances based on some gold standard data? I would like to use the same data to check the efficiency of my new approach.

It's hard for me to believe that there wouldn't be any such dataset, given that so many tools provide this service.

Any help would be appreciated.

[lieven.sterck]

did you check the papers describing those tools? perhaps they mention some sort of dataset?

[c_u]

In the papers for these tools, they do not report any dataset, just the fact that they provide these services. The papers of MrBayes has only the mathematical model used, the papers for MEGA provide performances but not for ancestral reconstruction, and I was unable to find the corresponding paper for Mesquite

Answer 1

[Jean-Karim Heriche]

For this, I would use simulation, see for example Computational reconstruction of ancestral DNA sequences.
If you want experimental data, have a look at An experimental phylogeny to benchmark ancestral sequence reconstruction.

[c_u]

Thanks Jean-Karim, the second paper was exactly what I was looking for. BTW, why do you think that using simulation is a better idea?

[Jean-Karim Heriche]

Because it's the only way to have absolute ground truth data. With simulations, you control and understand what's going on and so you can check sensitivity to different things by playing with parameters. I would use simulations in addition to the experimental data.

[c_u]

Got it, thanks a lot!