Chapter 14: Microbiomics: Analyzing Microbial Communities¶
14.1 We Are Not Alone¶
You are more bacteria than human. The Microbiome refers to the trillions of microorganisms living in and on us (and in the environment). These communities affect digestion, immunity, and even mental health.
Microbiomics (or Metagenomics) is the study of these communities.
14.2 16S rRNA: The Barcode of Bacteria¶
To identify which bacteria are present, we don't sequence their whole genomes. We sequence one specific gene: 16S rRNA.
- Universal: All bacteria have it.
- Variable: It has specific regions (V3, V4) that are different between species.
By sequencing this "barcode," we can identify the bacteria without growing them in a lab.
14.3 The QIIME 2 Workflow¶
QIIME 2 (Quantitative Insights Into Microbial Ecology) is the industry-standard command-line platform for this analysis.
- Import Data: Load FASTQ files.
- Denoising (DADA2): Correct sequencing errors and group identical sequences into ASVs (Amplicon Sequence Variants).
- Taxonomy Assignment: Compare ASVs to a database (like Silva or Greengenes) to give them names (e.g., E. coli).
- Diversity Analysis: Calculate how diverse the samples are.
14.4 Diversity: Alpha vs. Beta¶
13.4 Modern Microbiome Methods and Best Practices¶
Microbiome research can use targeted marker gene sequencing (16S) or whole-metagenome shotgun sequencing. Key practices:
- DADA2 / Deblur: Use amplicon denoising (DADA2) rather than OTU clustering for higher resolution.
- QIIME2: A full ecosystem for 16S processing, visualization, and reproducible pipelines.
- Shotgun metagenomics: Provides species- and strain-level resolution and functional profiling (e.g., HUMAnN).
Analysis considerations:
- Contamination control: Include blanks and negative controls; use
decontamto identify contaminants. - Compositional data: Use compositional-aware methods (ALDEx2, DESeq2 with caution) and transform data (CLR) before multivariate analyses.
- Diversity metrics: Alpha (within-sample) and Beta (between-sample) are standard; use rarefaction carefully and prefer robust normalization methods.
Data sharing:
- Deposit raw reads in SRA/ENA and provide sample metadata in standard formats (MIxS). Provide analysis notebooks and environment files for reproducibility.
This is the most important concept in microbiome analysis.
Alpha Diversity (Within Sample)¶
"How many different species are in this sample?" * Richness: Count of species. * Shannon Entropy: Accounts for both richness and evenness (abundance).
Beta Diversity (Between Samples)¶
"How different is the community in Sample A compared to Sample B?" * Jaccard Index: Based on presence/absence. * Bray-Curtis: Based on abundance. * UniFrac: Based on the phylogenetic distance between bacteria.
14.5 Bioinformatics in Action: Calculating Alpha Diversity¶
While QIIME 2 does this on a massive scale, let's write a Python function to understand the math behind Shannon Entropy, a common Alpha Diversity metric.
Where \(p_i\) is the proportion of the total population made up of species \(i\).
import math
def calculate_shannon_entropy(counts):
"""Calculates Shannon Entropy (Alpha Diversity) for a list of species counts."""
total = sum(counts)
if total == 0:
return 0.0
entropy = 0.0
for count in counts:
if count > 0:
# Calculate proportion (p)
p = count / total
# Add to entropy sum
entropy -= p * math.log2(p)
return entropy
# Example:
# Sample A: Very diverse (even spread)
sample_A = [10, 10, 10, 10]
# Sample B: Low diversity (dominated by one species)
sample_B = [37, 1, 1, 1]
entropy_A = calculate_shannon_entropy(sample_A)
entropy_B = calculate_shannon_entropy(sample_B)
print(f"Sample A (Diverse): {entropy_A:.2f}")
print(f"Sample B (Dominated): {entropy_B:.2f}")
Output:
Higher entropy means higher diversity!
Summary¶
Microbiomics uses the 16S rRNA gene to identify bacteria. We use tools like QIIME 2 to process reads into ASVs. We then compare communities using Alpha Diversity (richness within a sample) and Beta Diversity (differences between samples).