Alpha Diversity

An information-theoretic metric that jointly accounts for both richness and evenness.

• Formula: H' = -Σ pᵢ × ln(pᵢ), where pᵢ is the proportional abundance of taxon i
• Interpretation: Higher values indicate greater diversity. A value of 0 means a single taxon dominates.
• Sensitivity: Weights rare and abundant taxa more equally than the Simpson Index.
• Typical range: 0 to ~4.5 in most microbial communities.

Measures the probability that two randomly selected individuals belong to the same taxon, heavily weighting dominant taxa.

• Formula: D = Σ pᵢ² (Simpson's Dominance) or 1 - D (Gini-Simpson)
• Interpretation: A value near 0 (or 1 for Gini-Simpson) indicates high evenness.
• Key distinction: Less sensitive to rare species than Shannon; driven by the most abundant organisms.
• Use case: Preferred when the focus is on community dominance rather than total richness.

A non-parametric estimator of true total species richness, correcting for the fact that rare taxa are often unobserved in finite sequencing depth.

• Mechanism: Extrapolates total richness based on the ratio of singletons (taxa observed exactly once) to doubletons (taxa observed exactly twice).
• Formula: S_chao1 = S_obs + (a² / 2b), where a = singletons, b = doubletons.
• Critical caveat: Assumes a closed, well-mixed community and is sensitive to sequencing error inflating singleton counts.

Quantifies diversity not by taxonomic counts, but by the total branch length of the phylogenetic tree spanned by the taxa in a sample.

• Calculation: Sum of all branch lengths connecting the root to the observed taxa in a reference phylogeny.
• Advantage: Captures evolutionary divergence. A community with distantly related species scores higher than one with closely related species, even if richness is identical.
• Implementation: Computed using tools like QIIME 2 with a rooted phylogenetic tree (e.g., from Greengenes or SILVA).

The simplest alpha diversity metric: a direct count of the number of distinct Amplicon Sequence Variants (ASVs) or Operational Taxonomic Units (OTUs) detected in a sample.

• Strengths: Intuitive, no distributional assumptions.
• Weaknesses: Highly dependent on sequencing depth. Deeper sequencing will almost always detect more rare taxa, making comparisons across unevenly sequenced samples invalid.
• Mitigation: Must be used with rarefaction—randomly subsampling all samples to an equal sequencing depth before comparison.

A normalized measure of how equally abundant the taxa in a sample are, derived from the Shannon Index.

• Formula: J' = H' / ln(S), where H' is the Shannon Index and S is the total number of observed taxa.
• Range: Constrained between 0 (maximum unevenness, one taxon dominates) and 1 (perfect evenness, all taxa equally abundant).
• Utility: Decouples evenness from richness. Two samples can have identical Shannon values but different evenness if richness differs.

A foundational alpha diversity metric that quantifies the uncertainty in predicting the species identity of a randomly selected individual. It accounts for both richness (number of species) and evenness (relative abundance distribution).

• Calculated as H' = -Σ (p_i * ln p_i), where p_i is the proportion of species i
• Higher values indicate greater diversity
• Sensitive to rare species; widely used in microbial ecology

A non-parametric richness estimator that predicts the true total number of species in a sample by correcting for unseen rare taxa. It uses the ratio of singletons (species observed once) to doubletons (species observed twice).

• Essential when sampling depth is insufficient to capture full diversity
• Formula: S_est = S_obs + (a² / 2b), where a = singletons, b = doubletons
• Assumes rare species carry the most information about missing taxa

A dominance-weighted metric that measures the probability that two randomly selected individuals belong to the same species. It gives more weight to abundant species than rare ones, making it complementary to the Shannon Index.

• Calculated as D = Σ (p_i²), with the inverse (1/D) often reported
• Ranges from 1 (no diversity) to S (maximum evenness)
• Less sensitive to sampling effort than richness-based metrics

A visualization technique that plots the number of observed species against sequencing depth by repeatedly subsampling reads. It enables fair comparison of alpha diversity between samples of unequal sequencing effort.

• A plateauing curve suggests sufficient sampling depth
• Non-overlapping confidence intervals indicate significant differences
• Essential quality control step before downstream diversity analysis

An alpha diversity metric that incorporates evolutionary relationships by summing the branch lengths of a phylogenetic tree connecting all species in a sample. It captures the total evolutionary history represented.

• Requires a rooted phylogenetic tree of observed taxa
• A community of distantly related species yields higher PD than closely related ones
• Particularly relevant for functional diversity inference in metagenomics

The counterpart to alpha diversity that measures between-sample differences in community composition. It quantifies species turnover or nestedness across environmental gradients or treatment groups.

• Common metrics: Bray-Curtis dissimilarity, UniFrac, Jaccard index
• Visualized via Principal Coordinates Analysis (PCoA) or NMDS ordination
• Alpha and beta diversity together provide a complete picture of community structure