I found the same problem as other people said when readtree of a FastTree generated tree file.Do you mean as the tree generated by FastTree can’t be analyzed furthur by unifrac? Am I right?
I tried to construct a tree through Clearcut,as you recommended, however I don’t quite sure if it is for my too large data (about 10,000 reads) that it indicated:Clearcut: Memory allocation error in NJ_compute_dmat().
I think you have more experience and hope your help!
I guess I need to break down and actually try fasttree myself ;). I’m sure there’s a way to use fasttree-generated trees, but we don’t seem to have it right now. I think it’s because there are multifurcations and/or node labels, which will cause problems for us. As for clearcut, you might check with James Foster who regularly trolls around these parts and helped to develop clearcut.
I’m surprised to hear that clearcut blew up with “only” 10K sequences. How much memory does your machine have?
The students who wrote clearcut are all "Doctor"s now, and have moved on to greener pastures. So it is unlikely that we (meaning “I”) will have a more memory efficient implementation anytime soon.
My advice would be to use this as an excuse to buy more memory!
Yes,I found I made a mistake,not 10K reads but 100K reads.I am sorry!
How do you think of that? Does that possible? Actually it does not work in my computer. And furthur I’m going to process much larger data, what’s your suggestions?
how much memory do you have? with 100K I think you’re pushing the limits of any algorithms. Please post here if you find something that works!