Skip to content

Home

Jump to bottom
Danielle M. Stevens edited this page Aug 6, 2025 · 5 revisions

Welcome to the mamp-ml wiki!

This wiki currently contains troubleshooting advice if you run into any issues install mamp-ml and its many dependencies. Keep in mind this is not a comprehensive list and so if you run into an issue that can not be solved, please feel free to submit an issue and provide as much documentation as you can:

Useful information can include:

  • what computer/operating system are you running on
  • if you have a GPU available and if so, what mode
  • are you using an HPC
  • Output of the error

Guide to installing conda locally

If you have yet to install conda and are working on your own machine (not HPC), please use the directions below to help with install.

# go to this website https://www.anaconda.com/distribution/
# select the 3rd distribution if you don't already use anaconda 
# my computer is a linux, so I have downloaded the linux distribution 
# and can use bash to intall it
bash Anaconda-latest-Linux-x86_64.sh

# run so your computer can find conda with your path
# for anaconda 2 :
export PATH=~/anaconda2/bin:$PATH

# for anaconda 3 :
export PATH=~/anaconda3/bin:$PATH

# if you need to initalize conda
conda init 
source ~/.bashrc

Potential modifications for HPC users

Due to a combination of permission and security limitations, you may not be able to install mamp-ml directly using the install script. This is, in part, many HPCs have some software already available that you don't need to install. If this is you, please try installing manually rather than running the install_script with the following commands:

# if your HPC uses module to install software, please use that to activate conda
# example from HPC user #1
module load anaconda3

# example from HPC user #2
module load conda/latest

This will likely need to be modified depending on your HPC's set up, so please check their documentation for specifics.

Error: Can't activate/find conda environment

Depending on the computer I tested on, sometimes the conda environment names would get relabeled. If you have issues with this, check what condo environments you have installed and adjust the activation command accordingly.

# to list all the conda environments
conda info --envs

Which gave (for example) the following output:

/content/mamp-ml# conda info --envs
                                                                                
# conda environments:                                                           
#                                                                               
                       /content/mamp-ml/esmfold-conda/esmfold-conda             
                       /content/mamp-ml/localcolabfold/colabfold-conda          
                       /content/mamp-ml/localcolabfold/conda                    
base                   /usr/local

Then adjust the following run instructions:

# old 
# activate AlphaFold and run (please update your path)
conda activate localfold
export PATH="/global/scratch/users/dmstev/localcolabfold/colabfold-conda/bin:$PATH" 
colabfold_batch --num-models 1 ./mamp-ml/intermediate_files/receptor_full_length.fasta \
   ./mamp-ml/intermediate_files/receptor_only/

# new environment path
conda activate /content/mamp-ml/localcolabfold/colabfold-conda
export PATH="/content/mamp-ml/localcolabfold/colabfold-conda/bin:$PATH" 
colabfold_batch --num-models 1 ./mamp-ml/intermediate_files/receptor_full_length.fasta \
   ./mamp-ml/intermediate_files/receptor_only/

Clone this wiki locally