Singularity命令无法识别fastq.gz文件的问题求助

API 探路者

2026-6-26

问题描述

我正在用Snakemake和Singularity开发鸟枪法宏基因组分析流程，但在Snakemake中执行Singularity命令分析fastq.gz文件时遇到问题。单独运行Singularity命令后发现程序无法识别目标文件。我的Snakefile位于另一个目录，将数据复制到该目录后流程可正常运行，不清楚为何Singularity命令在数据原目录下无法工作。

执行的Singularity命令

singularity exec -B /mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/results/fastqc:/output docker://staphb/fastqc fastqc -o /output /mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R1_001.fastq.gz /mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R2_001.fastq.gz

命令输出

Using cached SIF image
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
        LANGUAGE = (unset),
        LC_ALL = (unset),
        LANG = "pt_BR.UTF-8"
    are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
Skipping '/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R1_001.fastq.gz' which didn't exist, or couldn't be read
Skipping '/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R2_001.fastq.gz' which didn't exist, or couldn't be read

文件权限信息

drwxr-xr-x 2 catg catg    0 abr  1 12:26 results
-r-xr-xr-x 1 catg catg 608M jan 13 16:06 Undetermined_S0_L001_R1_001.fastq.gz
-r-xr-xr-x 1 catg catg 606M jan 13 16:08 Undetermined_S0_L001_R2_001.fastq.gz

解决方案

挂载完整数据目录：当前仅挂载了输出目录，fastq文件所在的父目录未被挂载到Singularity容器内，容器无法访问宿主机未挂载的路径，因此找不到文件。修改-B参数，将数据所在根目录挂载：

singularity exec -B /mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN:/metagen docker://staphb/fastqc fastqc -o /metagen/results/fastqc /metagen/Undetermined_S0_L001_R1_001.fastq.gz /metagen/Undetermined_S0_L001_R2_001.fastq.gz

Snakemake中简化路径配置：在Snakefile里用内置的singularity参数块处理容器映射，避免手动写长路径，示例：

rule fastqc:
    input:
        r1="/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R1_001.fastq.gz",
        r2="/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R2_001.fastq.gz"
    output:
        "/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/results/fastqc/Undetermined_S0_L001_R1_fastqc.html",
        "/mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/results/fastqc/Undetermined_S0_L001_R2_fastqc.html"
    singularity:
        "docker://staphb/fastqc"
    shell:
        "fastqc -o {output[0].dirname} {input.r1} {input.r2}"

Snakemake会自动处理容器内的路径映射，无需手动指定-B参数。

检查父目录权限：确认从/mnt/share到/METAGEN的每一层父目录都有执行权限（x权限），否则容器内无法遍历路径。可执行以下命令逐层检查：
```
namei -l /mnt/share/Equipamentos_NTO/Med_Pers/Illumina_NextSeq/Corridas/run142/BIOINFO/METAGEN/Undetermined_S0_L001_R1_001.fastq.gz
```

内容的提问来源于stack exchange，提问作者Rafaela Carolina dos Anjos Sch