1 Data Tidying

1.1 General snakemake workflow

A tentative snakemake workflow that defines data processing rules in a DAG (directed acyclic graph) format. A detailed interactive snakemake HTML report is available here. Use a wider screen to get a better interactive snakemake report.

1.2 Data Tidying

Requires:

• Sample metadata
• OTU tables from mothur and qiime2 pipelines.
• Taxonomy tables from mothur and qiime2 pipelines.

1.3 Microbiome composite R objects

Requires:

• Tidy sample metadata.
• Tidy OTU tables.
• Tidy taxonomy tables.

The mothur and qiime2 composite objects are in a long-format which is suitable for most types of data visualization.

1.4 Microbioma phyloseq objects

Requires:

• A phyloseq package
• Tidy sample metadata.
• Tidy OTU tables.
• Tidy taxonomy tables.

1.5 Data transformation

Requires:

• A phyloseq package
• A microbiome package
• Phyloseq objects

Data transformation is intended to converting the values into ready-to-use matrices. There are different methods out there, and here are just a few:

• No transformation is similar to raw abundance.
• Compositional version or relative abundance.
• Arc sine (asin) transformation.
• Z-transform for OTUs.
• Z-transform for Samples.
• Log10 transformation.
• Log10p transformation,
• CLR transformation.
• Shift the baseline.
• Data Scaling.

1.6 Data reduction

Dimensionality reduction is intended to reduce the dimension of the variables but keeping as much of the variation as possible. , and includes:

1. Linear methods (commonly used in microbiome data analysis)
   • PCA (Principal Component Analysis)
   • Factor Analysis
   • LDA (Linear Discriminant Analysis)
   • More here.
2. Non-linear methods