So hi! I can't access my other account since I can't login through facebook at the time so I created this other account. Hi Fellow scientists, I'm a grad student in Wildlife Science and I can't seem to wrap my head in building a phylogenetic tree involving multiple genera. It's not really an assignment but I am kinda falling behind in my class due to this. A help would be greatly appreciated! We had a similar thing in the lab class but this thing is similar to it as we try to list traits and determine their characters states across genera.

My conclusion here is that Nyctixalus is the closest to the outgroup; therefore, more basal than the rest. What I start to have trouble with was with Philautus, Chiromantis, Polypedates, and Rhacophorus. Philautus had 1 ancestral trait it shares with Platymantis (Direct Development), at the same time Polypedates and Chiromantis also share 1 ancestral trait which is Partial Webbing. But Polypedates and Chiromantis also share multiple derived traits with Rhacophorus. My fellow group also had came across the same problem when we were doing the activity like how do we construct this. This is my assumption:

But I realized that this make Rhacophorus related to Philautus at the same level as Chiromantis and Polypedates when Rhacophorus is more related to Chiromantis and Polypedates due to Foam Nests than it is to Philautus. How do I fix this? I tried consulting AI and it can't show me a proper construction of the tree. A help would be gladly appreciated.

A few quick thoughts without going into too much depth due to time constraints:

Just generally speaking, it is important to know how distances are scored and if they are weighted. Generally speaking you calculate the distances for each taxa resulting in a distance matrix and then build the tree using methods such as Neighbor-Joining, or, what is seemingly applicable here, UPGMA (but I suspect that it is not really in the assignment).

It may have been discussed in class but looking at the provided example, most are binary and I assume that there is no special scoring going on. But webbing has three descriptors and they could either be equidistant, i.e. the score from full to absent is the same as full to partial, or, perhaps more likely, the distance would absent -> partial -> present. There, you could either score the difference from absent to present the same as in the other categories (e.g. 0 for same, 1 for different) and have partial in-between (e.g., 0.5). Or you could score each jump fully (e.g., 0, 1, 2), meaning that from absent to present the distance would be higher than in any of the other binary categories.

To provide some examples:

The distance between Rhacophorus to Polypedates could then either 0.5+0+0+0 = 0.5 or it could be 1 +0+0 +0=1.

And you would continue to do that for every pair. In the resulting distance matrix you look for the shortest distance and join them. Normally, we would calculate branch lengths for both that are equidistant to the node where they are connected. In this case, you cannot really do that, as the distance is 0 as all listed parameters between Chiromantis and Polypedates are identical. I.e. just using this information you wouldn't put them into different branches in the first place.

What you would normally do then is calculate a new distance matrix which is reduced in size. In the above example, Chiromantis and Polypedates would form one cluster and the difference from that one two all others would be calculated in the updated matrix.

In OP the assignment is simplified to be able to skip these detailed steps, but the basic idea is still to calculate the distances and build from there (and ignoring branch lengths. But in a proper UPGMA method, the distance would move nodes at different depths.

Thanks for the reply, that is really insightful as there is more to it that I didn't know. I actually had a hard time understanding this distance matrix, but our class is just introductory to phylogeny, so I don't think the distances of the branches doesn't matter that much here? What my prof wants us to know is how to relate them based on traits I assume.

So Chiromantis and Polypedates would form one cluster, I get that but where do I place Rhacophorus? Should it branch from the Chiromantis and Polypedates clade? Should it have it's own branch? It's actually the one with the most derived traits. So if our prof decided not to combine Chiromantis and Polypedates should I just make this alongside Rhacophorous into a Polytomy?

It would greatly help me to know this to avoid similar situations if they arise again. Thank you!

9 hours ago, CharonY said:

A few quick thoughts without going into too much depth due to time constraints:

Just generally speaking, it is important to know how distances are scored and if they are weighted. Generally speaking you calculate the distances for each taxa resulting in a distance matrix and then build the tree using methods such as Neighbor-Joining, or, what is seemingly applicable here, UPGMA (but I suspect that it is not really in the assignment).

It may have been discussed in class but looking at the provided example, most are binary and I assume that there is no special scoring going on. But webbing has three descriptors and they could either be equidistant, i.e. the score from full to absent is the same as full to partial, or, perhaps more likely, the distance would absent -> partial -> present. There, you could either score the difference from absent to present the same as in the other categories (e.g. 0 for same, 1 for different) and have partial in-between (e.g., 0.5). Or you could score each jump fully (e.g., 0, 1, 2), meaning that from absent to present the distance would be higher than in any of the other binary categories.

To provide some examples:

The distance between Rhacophorus to Polypedates could then either 0.5+0+0+0 = 0.5 or it could be 1 +0+0 +0=1.

And you would continue to do that for every pair. In the resulting distance matrix you look for the shortest distance and join them. Normally, we would calculate branch lengths for both that are equidistant to the node where they are connected. In this case, you cannot really do that, as the distance is 0 as all listed parameters between Chiromantis and Polypedates are identical. I.e. just using this information you wouldn't put them into different branches in the first place.

What you would normally do then is calculate a new distance matrix which is reduced in size. In the above example, Chiromantis and Polypedates would form one cluster and the difference from that one two all others would be calculated in the updated matrix.

In OP the assignment is simplified to be able to skip these detailed steps, but the basic idea is still to calculate the distances and build from there (and ignoring branch lengths. But in a proper UPGMA method, the distance would move nodes at different depths.

Thank you so much for this information.

I had no idea the life scientists had advanced so far, mathematically, but it makes perfect sense.

Mathematicians have been busy expanding the notions of measurements and distance so I am glad to see that bearing fruit in the form of real applications.

+1

15 hours ago, Frenzypopcorn said:

Thanks for the reply, that is really insightful as there is more to it that I didn't know. I actually had a hard time understanding this distance matrix, but our class is just introductory to phylogeny, so I don't think the distances of the branches doesn't matter that much here? What my prof wants us to know is how to relate them based on traits I assume.

Even if you are not concerned about the precise distance, the idea here is to formalize what is closer and what is more distant to each other. This is where the distance matrix is important. What you would do create a table. For example for 3 genera it would look like this.

1 2 3

1

2

3

And for each you calculate the distance (or difference) between them. 1:1 would be 0 for example, as it would be the same genus. So if you follow the list in your table (1 Rhacophorus, 2 Polypedates, etc.). You then would calculate 1:2 -> 0.5 (assuming the partial to full is scored as 0.5) 1:3 -> 0.5, 2:3 -> 0 (that is what makes it a bit iffy as the distance between different genera should be >0).

So you would build up the first node from the smallest distance (0, with all its iffyness), so a common node connecting 2 and 3. 1 is equidistant to both, so a new node would connect 1 to 2:3. What you would do is do the same calculations for every pair and then build your tree from there. I think it helps if you integrate nodes into your trees (which are the branching points). You can rotate trees around them without changing branch distances.

For example the following trees are all depicting the same relationship, they are just rotated:

8 hours ago, studiot said:

Thank you so much for this information.

I had no idea the life scientists had advanced so far, mathematically, but it makes perfect sense.

Mathematicians have been busy expanding the notions of measurements and distance so I am glad to see that bearing fruit in the form of real applications.

I would say it is more advanced (or perhaps just easier) in the genetic field as the mathematical models are clearer there, as we basically calculate distance based on base differences. This allows other models (such as Maximum Likelihood). If you are interested Nei and Tamura have done a lot of work on it in the 90s. And have also be foundational in the area of molecular clocks.

I am really lost here, is it okay to request for a visual example if you don't mind? The ratios confuses me a lot.

@CharonY

Perhaps a moderator would help you access/combine your old account as I can see your old posts under the user 'popcornfrenzy' ?

With regards to this question here is a free nifty visual network tool that allows you to draw and calculate the distance (weighted or unweighted) matrix amongst other things.

It has an instructional video.

https://graphonline.top/en/

Have fun drawing your tree and calculating your distances.

On 10/9/2025 at 1:27 AM, Frenzypopcorn said:

I am really lost here, is it okay to request for a visual example if you don't mind? The ratios confuses me a lot.

Also look here for some worked examples

Biology LibreTexts

26.3: Distance Based Methods

On 10/8/2025 at 6:27 PM, Frenzypopcorn said:

I am really lost here, is it okay to request for a visual example if you don't mind? The ratios confuses me a lot.

There are no ratios. 1:2 refers to the distance between the first genus on your table compared to the second. So, Racophorus : Polypedates. If you look at the listed traits all are identical except webbing. So the difference except webbing is 0 (i.e. no distance/identical).

The difference in webbing is "full" vs "partial". Here it boils down how you evaluate that difference. If you think that full vs partial is the same as full vs absent, then the score would be the same (e.g., 1). If you think full vs partial is less of a difference, then you could e.g. score it with a 0.5.

You would follow the approach for all possible combinations (hence, the matrix shape).