Alpha and Beta Diversity analyses fail in Statistical Meta-Analysis module (16S soil metabarcoding data)

MicrobiomeAnalyst
, ,
1

16S Tax.csv (780.6 KB)
16S-Black.csv (204.4 KB)
16S-White.csv (204.3 KB)
meta1.csv (324 Bytes)

Hi, I’m currently working with 16S rRNA metabarcoding data derived from soil surface samples, and I would like to ask for your support regarding an issue I’m encountering in the Statistical Meta-Analysis module.

My study compares two biological conditions:

  • “Black” samples, which come from soil plots with high visible microbial growth, and
  • “White” samples**, which come from corresponding plots with low or no visible microbial activity*.

There are a total of 22 samples, consisting of 11 paired sites — each site includes one black and one white sample from the same physical location.

For the analysis, I uploaded two OTU/ASV abundance tables:

  • 16S-Black.csv (ASV table for black samples)
    *- 16S-White.csv (ASV table for white samples)
    In addition, I used a shared taxonomy file (16S-Tax.csv) and a shared metadata file (meta1.csv), which includes a Location column indicating whether a sample is black or white. All sample names match across the files and have no missing values.

The data upload and processing steps run smoothly — both datasets are successfully recognized, and the “Proceed” function works without any issues. I am also able to use the Visual Exploration features (e.g. PCoA, stacked bar plots), and grouping by Location is correctly recognized.

However, I run into problems when I try to calculate diversity indices:

  • When selecting Alpha Diversity, I am immediately redirected to the troubleshooting page, with no error message or indication of what went wrong.
  • In Beta Diversity, I can access the settings page and select parameters such as ordination method and distance metric. However, once I click “Update”, the analysis fails and I receive only the following generic error message: “Possible causes of error (last one being the most relevant):”
    Unfortunately, no further explanation is provided.

I’ve double-checked my metadata and grouping structure, and everything seems to be in order. Given that the visualizations and initial processing steps work perfectly, I am unsure what is preventing the diversity analyses from running.

Do you have any insight into what might be causing this issue?
I would be happy to provide example files if needed.

Thank you very much in advance for your time and support!

Warm regards,
Lina

2

Thanks for the detailed info. Without testing your data (due to limited time), some quick comments based on your descriptions

  1. Meta-analysis is usually after exploring individual data. Have you tried to perform individual data analysis in the Marker Data Profiling module?
  2. Meta-analysis is usually for multiple datasets (minimal 3) that are expected to share a lot of common features (50%?), so that you combine more samples to increase statistical power (i.e. more samples) for evaluating changes of these features. If you only have two data, you can directly merge them into one and add one factor such as “Study” as a new metadata column for Marker Data Profiling
3

Thanks a lot. It worked with one dataset, including all sample points, in the feature Marker Data Profiling, not Statistical Meta-Analysis.