udem-lbit / coretracker Goto Github PK

Hi!

I am interested to use the CoreTracker program, but encountered some problems during the installation.
I have followed the instruction for conda installation but was given this error message:

Fetching package metadata: ..........
Solving package specifications: ....

The following specifications were found to be in conflict:
  - coretracker
  - mkl_fft (target=mkl_fft-1.0.15-py27ha843d7b_0.tar.bz2) -> numpy >=1.11.3,<2.0a0
  - mkl_random (target=mkl_random-1.1.0-py27hd6b4f25_0.tar.bz2) -> numpy >=1.14.6,<2.0a0
Use "conda info <package>" to see the dependencies for each package.

I have tried to reinstall the numpy and scipy as recommended, but the problem still persist.
Do you know how to troubleshoot this?

Thanks!

Hi @maclandrol,

Sorry to bother you again.
I am having another problem with the CoreTracker.
The program can run but it seems to stop at a certain point when predicting the possible codon reassignment. The run has never finished even after a few days.
The only error message that I received was:

ETE: cannot connect to X server

Do you know how to resolve this issue?

Many thanks

muscle, mafft and hmmer package are not checked during installation. Download and install them directly with setup.py

Cannot install using pip without -e.

Hi @maclandrol,
I am having an issue while running CoreTracker.
I am getting the program to run successfully, however, I noticed that it consistently filters out the sequences of interest to me. I input sequences from at least 70 different species, but the only ones that appear to be missing from the output codon_data files are the ones that I care the most about. While running, CoreTracker tells me "DEBUG:root:Non-uniform sequence in sequences and tree. The tree will be pruned." So I'm assuming that my sequences of interest are pruned out. I would like to know what constitutes "non-uniform" and if there is a way that I can keep them from being removed? I'm aware that those sequences are particularly divergent from the rest...which is exactly why I want to use CoreTracker on them!

Thank you for your help,
Nicole