can't plot either Consensus Tree or NLDR plot

[drelo]

Dear Huang et al,

I am following the tutorial example and when I try to plot the consensus tree nothing appears and when I try to plot the NLDR only the image below appears, but the button Plot Tree doesn't work. All the NLDR results I obtained are in the attached zip.

(I am using Ubuntu 16.04 I can provide additional details as requested, please let me know how can I fix this. Thanks)

results.zip

[drelo]

Hi again, here are the files obtained before the NLDR
Thanks for your help

beforeR.zip

[whuang08]

Hi,
I just fixed the problem about plotting a consensus tree. The new version should work.
I did not see problems when I plot a 2D NLDR result. Did you check the "weighted trees" box when you load trees? If not, then all the edges in all the trees have weights 1, which is clear not what we want. It follows that the distances between the same topology trees are 0. Therefore, the NLDR plot will only have a few points.

[drelo]

Yes, I used 'weighted trees'
I just downloaded and tried the new version the consensus plot now works fine but the NLDR doesn't work it just return an image similar to the previous one.

Attaching the files

LAST.zip

[whuang08]

Can you tell me all the exact step to plot an NLDR picture? I just tried again and still don't see the problem. I would like to reproduce the problem so that I can debug. Thanks.

[drelo]

I loaded the tress, computed bipartition matrix, compute consensus, plot it, compute T to T distances, go to the next tab, select dimension 3 and step size 100, click Run NLDR , pick a folder ( I choose one containing the Linux... extracted folder, called TREESCAPER), and the click plot results which shows this
Sorry if this sounds confusing, probably I am doing something wrong.

[whuang08]

So you change the step size in the "plot parameter", right? So what is the number of cluster?
My setting is:

[drelo]

Sorry to bother you again, I did it without the 'step' setting, cluster is 1, I checked and I have everything as in your pic of the dialog. It still gives me the same image.

[whuang08]

Can you also send me the weighted RF-distance file? The following is the NLDR parameters that I used, which are also the default values.

[drelo]

Sure, Here is the weighted RF distance file!

Linux_TreeScaper_v1_unrooted_weighted_RF-distance.txt

[whuang08]

En. It is a zero matrix, which does not make sense.

[drelo]

But the raw tees are the ones in the example file, there might be something odd in the calculation of the distances. I just used another machine and did a new installation from scratch and I got the same results (the figure with only 1 point)

Here are the results loading the pre computed distances (unweigthed RF) from the tutorial files Output folder. But

[whuang08]

I agree with that there may be something odd in the distances. However, I am not able to reproduce the bug in my computer. I can correctly compute weighted RF distance matrix. I will find a new Ubuntu computer and see if I can reproduce the bug.

[drelo]

Thanks for your help! let me know if I can assist by sending or processing any file.

[whuang08]

Weird. I find a new computer which uses Ubuntu 16.04.01. It is the same as the one you have, isn't it?
I still can not reproduce the problem :(
BTW, did you use Unweighted RF-distance or Weighted distance? The picture in your previous message seems to be the result by Unweighted RF-distance.

[drelo]

Yes, I have 16.04.

I loaded the unweighted distances from the

Following the tutorial I tried weighted RF

Here is a GIF of what I did