Let's tackle this Bioconda channel 404 issue you're facing. The error means your build process can't reach the official Bioconda channel, even after adding it to your conda config. Here are actionable fixes to resolve this:

1. Use a Specific, Stable Miniconda Version #

The latest tag for Miniconda can sometimes pull an outdated version that has compatibility issues with newer channel structures or HTTPS endpoints. Instead, specify a known-stable version that matches your Python requirement:

ARG CONDA_VER=py39_4.12.0  # Stable version compatible with Python 3.9

Also update the download URL (Continuum Analytics is now part of Anaconda, so the repo URL has changed):

RUN curl -LO "https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VER}-Linux-${OS_TYPE}.sh"

2. Adjust Channel Priority & Order #

Bioconda recommends strict channel priority and a specific channel order to avoid conflicts. Add this config step before installing packages:

RUN conda config --set channel_priority strict

Reorder your channel additions to follow the recommended sequence (conda-forge first, then bioconda, then defaults):

RUN conda config --add channels conda-forge && \
    conda config --add channels bioconda && \
    conda config --add channels defaults

3. Switch to a Bioconda Mirror (if network is the issue) #

If the official channel is unreachable due to network restrictions or outages, use a trusted mirror. For example, the Tsinghua University mirror:

RUN conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/ && \
    conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/bioconda/ && \
    conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/

4. Verify Network Access in Docker Build #

Sometimes Docker's default network setup blocks external requests. Try running the build with host network mode to rule out isolation issues:

docker build --network host -t your-image-name .

Modified Dockerfile Example #

Here's your Dockerfile with all fixes applied:

ARG UBUNTU_VER=18.04
ARG CONDA_VER=py39_4.12.0  # Specific stable version
ARG OS_TYPE=x86_64
ARG PY_VER=3.9
ARG PANDAS_VER=1.3
ARG BEDTOOLS_VER=2.3

FROM ubuntu:${UBUNTU_VER}

# Copy clinvar and parse_vcf.py to directory called /mydir/
COPY . /mydir/

# System packages
RUN apt-get update && apt-get install -yq curl wget jq vim

SHELL ["/bin/bash", "-c"]

# Use the above args
ARG CONDA_VER
ARG OS_TYPE

# Install miniconda to /miniconda with updated URL
RUN curl -LO "https://repo.anaconda.com/miniconda/Miniconda3-${CONDA_VER}-Linux-${OS_TYPE}.sh"
RUN bash Miniconda3-${CONDA_VER}-Linux-${OS_TYPE}.sh -p /miniconda -b
RUN rm Miniconda3-${CONDA_VER}-Linux-${OS_TYPE}.sh

ENV PATH=/miniconda/bin:${PATH}
RUN conda update -y conda
RUN conda init

ARG PY_VER
ARG PANDAS_VER
ARG BEDTOOLS_VER

# Configure conda channels with correct priority and order
RUN conda config --set channel_priority strict && \
    conda config --add channels conda-forge && \
    conda config --add channels bioconda && \
    conda config --add channels defaults

# Install packages with specified versions
RUN conda install -y python=${PY_VER} pandas=${PANDAS_VER} bedtools=${BEDTOOLS_VER}

RUN conda config --show channels
RUN conda config --show-sources

Why This Works #

• Specific Miniconda Version: Avoids bugs in outdated latest builds and ensures compatibility with modern channel endpoints.
• Channel Priority: Prevents default channels from overriding bioconda/conda-forge packages, which can cause resolution failures.
• Mirror Channels: Provides an alternative reachable endpoint if the official channel is blocked.
• Host Network: Eliminates Docker network isolation as a potential cause of the 404 error.

内容的提问来源于stack exchange，提问作者moth