cluster summary file will show incorrect values for numerical entries when loaded in Excel

[apetkau]

Description of the bug

When running arboratornf on data that includes numerical metadata values, the cluster_summary.tsv value should summarize values from different samples in a single cell for each grouped set of samples (e.g., as a list of values or min/max).

These values are properly stored in the TSV tabular file as comma-separated numbers (e.g., 1000,2000). However, in Excel, these are incorrectly interpreted as a single number with misplaced commas and will be improperly shown (e.g., as 10,002,000).

For example, given this set of metadata as input:

Portion of samplesheet.csv with some columns renamed to better match output:

sample  outbreak      value
S1      1             1000
S2      1             2000
S3      2             3000
S4      2             4000
S5      3             5000
S6      unassociated  6000

The cluster summary for the metadata_partition group 1 will be:

Portion of cluster_summary.tsv

Outbreak Code  value
1              1000,2000
2              3000,4000
3              5000
unassociated   6000

But, when loading cluster_summary.tsv in Excel, the following is shown:

cluster_summary.tsv in Excel

Command used and terminal output

To reproduce this behaviour, please run the following with the attached samplesheet.tsv

nextflow run . -profile test,docker --outdir results --input samplesheet.csv --metadata_1_header value

To verify the output as a text file and in Excel, please download the attached output cluster summary file.

Relevant files

• Input samplesheet for pipeline: samplesheet.csv
• Output cluster summary from pipeline (inside zip file): cluster_summary.zip

System information

• Nextflow version: 24.04.4
• Hardware: Desktop
• Executor: local
• Container engine: Docker
• OS: WSL/Ubuntu
• Version of arboratornf: 0.1.0