CentOS7下Conda安装Snakemake遇依赖冲突及JSON解析错误求助

Hey there, let's work through this snakemake installation headache step by step—conda dependency conflicts and JSON parsing errors are super common, but we've got solid fixes to try:

1. Resolve the JSON Parsing Error First #

That ValueError: Expecting : delimiter is almost always caused by corrupted conda cache files. Let's clear them out to start fresh:

conda clean --all -y

This command deletes all cached packages and index files, so conda will fetch uncorrupted, up-to-date data next time you run an install.

2. Create a Fresh Conda Environment (Most Reliable Fix) #

Your existing environment likely has tangled dependencies, and Python 3.4.5 is quite outdated—newer snakemake releases don't even support Python 3.4 anymore. Let's build a clean environment tailored for snakemake:

• Create a new environment with a snakemake-compatible Python version (3.6–3.9 work perfectly):

  conda create -n snakemake_env python=3.8 -y
  

• Activate the new environment (CentOS 7 may require source for older conda versions):

  source activate snakemake_env
  

• Install snakemake with properly ordered channels (conda-forge first to avoid dependency clashes):

  conda install -c conda-forge -c bioconda snakemake -y
  

3. Fix the Backports Dependency Conflict Directly #

If you absolutely need to use your existing environment, try forcing a compatible version of backports.functools_lru_cache that plays nice with snakemake:

conda install -c bioconda -c conda-forge snakemake backports.functools_lru_cache=1.5 -y

Version 1.5 is a known compatible release for older snakemake versions that support Python 3.4.

4. Configure Conda Channel Priority #

Misconfigured channel order often leads to pulling incompatible package versions. Set up your channels to prioritize reliability:

conda config --add channels conda-forge
conda config --add channels bioconda
conda config --add channels defaults
conda config --set channel_priority strict

This ensures conda picks packages from the highest-priority channel first, cutting down on cross-channel dependency conflicts.

A quick heads-up: Python 3.4 is end-of-life, so moving to a newer Python version (even just in a conda environment) will save you from tons of future dependency headaches.

内容的提问来源于stack exchange，提问作者Nikita Vlasenko