Sarah Lutteropp

Observations from this experiment: - NetRAX does similarly well as raxml-ng in inferring the correct tree. - On very few datasets, raxml-ng inferred one near-zero-branch in its ML tree.

Good question. I did not include this in the reported results file, but will check by hand by looking at the raw data (after asking Alexey what length is used...

Adding Alexeys Slack answer here for cross-reference: ``` it is "--blmin" which is 10^-6 by default: https://github.com/amkozlov/raxml-ng/blob/master/src/main.cpp#L2201 ```

 I installed Aliview and made a screenshot of what the simulated MSA looks like... turns out 100 sites provide not enough signal for 10 taxa. **Maybe...

Things people wrote about this issue in the slack already (I had another dataset there, but the issue is the same): benoit: if you want to increase the signal in...

Turns out the 100 MSA sites were a typo, the simulation setting we agreed upon had 500 or 1000 sites (https://github.com/lutteropp/NetRAX/issues/3#issuecomment-730342835)... But still, the same scaling issue is expected to...

This issue is very low priority. I am printing the number of near-zero branches in raxml-ng best ML tree now in the results CSV file, turns out this happens very...

Do we need to add these Dendroscope distances to the simulated network also for the maximum-likelihood tree inferred by raxml-ng? If so, how to solve the rooting issue, since the...

I decided to not compute those Dendroscope topological distances for the raxml-ng best tree. After all, our goal is to infer networks, not look at how much worse raxml-ng tree...

Adding the parameters used by Celines simulator to the CSV output. But keep in mind that the exact same choice of simulation parameters shows extreme variance in terms of number...