conda-environment-creation-and-management

conda-environment-creation-and-management

Summary

Create and activate isolated Conda environments from YAML specification files to deploy multi-language scientific pipelines (Python, R, compiled binaries) with pinned dependency versions. This skill ensures reproducible installation of Hi-C data processing tools and their dependencies across heterogeneous computing environments.

When to use

You need to deploy a complex multi-language pipeline (e.g., HiC-Pro) that requires Python >3.7 libraries (pysam, bx-python, numpy, scipy), R packages (ggplot2, RColorBrewer), and compiled tool binaries (bowtie2, samtools >=1.9) in a way that is reproducible, isolated from system packages, and portable across laptops and HPC clusters. Use this skill when an environment.yml specification file is provided and you want to avoid manual installation of interdependent tools with conflicting system dependencies.

When NOT to use

• The target environment already exists and is actively in use — use conda update or environment reconstruction instead of create.
• You require a container-based deployment (Docker or Singularity) — use HiC-Pro's pre-built Docker image from Docker Hub or Singularity recipe instead.
• Python or R dependencies have unresolvable conflicts in Conda — fall back to manual compilation or virtual environments (venv) for isolated Python-only workflows.

Inputs

• environment.yml specification file (YAML format with Python, R, and bioconda package declarations)
• Miniconda or Anaconda distribution (pre-installed or to be installed)
• Target installation path (directory where environment will be created)

Outputs

• Isolated Conda environment directory with all dependencies installed and verified
• Activated environment shell session ready for pipeline execution
• Summary report documenting all resolved tool paths, library versions, and dependency integrity

How to apply

First, install Miniconda if not already present by following the official Miniconda installation documentation. Second, create a Conda environment from the HiC-Pro environment.yml file using conda env create -f environment.yml -p <installation_path>, where the -p flag specifies the full path for the environment rather than installing into the default envs directory. Third, activate the environment using conda activate <installation_path>. Fourth, verify Python version is >3.7 and test import statements for critical libraries (bx-python >=0.8.8, numpy >=1.18.1, scipy >=1.4.1, pysam >=0.15.4). Fifth, verify R availability and test R package installation via library(ggplot2); library(RColorBrewer) calls. Finally, verify that tool binaries bowtie2 and samtools >=1.9 are in PATH and executable, noting that bowtie2 and samtools can be automatically installed by HiC-Pro if not detected, but iced must be independently installed from https://github.com/hiclib/iced since it is no longer bundled with HiC-Pro.