The Second Genome as an Evolutionary Inheritance System

Executive hypothesis

The microbiome can function as an extended phenotype: a host-associated, partially heritable biological system whose composition and activity alter host traits, fitness, and evolutionary trajectories. Under this hypothesis, natural selection does not act only on host alleles that build phenotypes directly. It can also favor host genotypes, behaviors, and life histories that acquire, maintain, transmit, or regulate microbial partners that improve performance in a given environment.

This claim is stronger than the observation that microbes influence hosts. It predicts that microbiome-mediated effects can become recurrent across generations, generate heritable variation in host fitness, and redirect host adaptation—even when the relevant causal molecules are encoded by microbial genomes rather than the host genome.

Why the hypothesis is plausible

Microbes alter digestion, immune development, pathogen resistance, hormone metabolism, thermal tolerance, and behavior. Hosts also influence microbial communities through gut anatomy, secretions, immune filtering, diet, social contact, and vertical transmission. These reciprocal effects create the basic conditions for an evolutionary feedback loop.

The microbiome is not uniformly inherited. Some members are transmitted through eggs, birth, milk, nests, or social contact; others are repeatedly acquired from the environment. This variability does not invalidate the hypothesis. It creates a gradient of evolutionary regimes. Highly faithful transmission can permit coadaptation between host lineages and microbial consortia. Low-fidelity acquisition can still produce selection if host traits consistently bias which environmental microbes are recruited or retained.

Several established phenomena make the proposal testable. Host genetic variants are associated with microbial composition; microbial metabolites can alter host physiology; and experimental changes in microbiota can affect host growth, immunity, and reproduction. The unresolved question is whether these effects persist as a population-level evolutionary force rather than remaining transient ecological influences.

Mechanistic sketch

Consider a host locus that modifies gut mucus chemistry. That locus may change the abundance of a bacterium producing a metabolite that improves nitrogen acquisition. The host genotype therefore affects a microbial phenotype, which affects host nutrition, survival, and fecundity. Selection can favor the host allele indirectly through its microbial consequence.

The reverse pathway is equally important. A microbial gene cluster may produce a compound that improves host pathogen resistance. If the microbe is transmitted with sufficient fidelity, microbial variants that increase host survival may spread because they are carried into more future host-associated habitats. Host and microbial fitness need not be perfectly aligned: conflict may arise over resource use, transmission routes, or immune suppression. Evolutionary outcomes should therefore range from mutualistic coadaptation to conditional exploitation.

Mathematically, the relevant unit is not necessarily an organism but a recurring association. The key parameters are transmission fidelity, host control over community assembly, effects on host fitness, microbial persistence, and the covariance between microbial state and the reproductive success of host lineages.

Falsifiable predictions

• Transmission prediction: microbiome-mediated host traits should show greater parent–offspring resemblance when transmission is experimentally enforced or naturally high.
• Selection prediction: host populations exposed for multiple generations to contrasting microbial environments should diverge in host traits, even when host genomes are initially identical or periodically homogenized.
• Recruitment prediction: host genotypes under selection should evolve altered microbial filtering—through immune, anatomical, dietary, or behavioral mechanisms—rather than merely altered responses to identical communities.
• Reciprocal adaptation prediction: microbes isolated from hosts adapted to a particular environment should confer larger fitness benefits to those host backgrounds than to foreign backgrounds, producing host-by-microbiome interactions.
• Disruption prediction: antibiotics, germ-free rearing, or microbiome transplantation should weaken or redirect evolutionary responses attributed to host adaptation.

Experimental roadmap

A decisive test would combine replicated experimental evolution with defined microbial communities. Begin with genetically matched animals—such as Drosophila, zebrafish, or nematodes—and inoculate them with microbial consortia differing in metabolic capacity but not initial cell density. Maintain host populations for many generations in contrasting environments, such as low nitrogen, pathogen exposure, or thermal stress.

At regular intervals, measure host survival, development, fecundity, behavior, immune phenotypes, microbial abundance, strain-level genomes, and host gene expression. Perform reciprocal microbiome transplants between evolved host lines and common-garden assays in which host and microbial effects are separated. Whole-genome sequencing can identify host allele-frequency change; metagenomics and metabolomics can identify microbial functions under selection.

A second experiment should manipulate transmission directly. Compare vertical, socially mediated, environmental, and randomized microbial inheritance. If the microbiome acts as an extended phenotype, increased transmission fidelity should strengthen multigenerational persistence of microbiome-associated host fitness differences, while randomization should reduce it—unless host recruitment mechanisms evolve.

Controls and pitfalls

Critical controls include sterile or antibiotic-treated hosts, sham inoculations, heat-killed microbial preparations, cross-fostering, cage effects, and repeated independent microbial isolates. Host density, diet, maternal condition, and pathogen exposure must be standardized because each can generate apparent microbiome inheritance without stable microbial causation.

The largest conceptual pitfall is confusing correlation with evolutionary inheritance. A microbial taxon may track host genotype without causing fitness differences. Another is treating community composition as the phenotype when function is the relevant variable: unrelated taxa may perform equivalent metabolic roles. Experiments should therefore prioritize strain-resolved functional genes, metabolites, and causal re-inoculation.

Low transmission fidelity may also prevent sustained coadaptation, while rapid microbial evolution can obscure host genetic responses. These are not failures of the hypothesis; they define boundary conditions. The proposal predicts that microbiome-mediated evolutionary effects should be strongest when functional microbes persist, transmit, and covary with host reproduction.

Implications

If supported, this framework would broaden the concept of heredity. Evolutionary change could proceed through host alleles, microbial genes, host-controlled ecological filters, and recurring associations among them. It would also alter how adaptation is reconstructed: unexplained host traits might reflect selection on microbial partnerships, while conservation strategies could require preserving microbial transmission networks rather than host genomes alone.

The hypothesis does not imply that every microbiome is an adaptive organ or that hosts and microbes form a single indivisible individual. It proposes a measurable possibility: under identifiable ecological and transmission conditions, microbial communities become stable contributors to host fitness and therefore participants in host evolution. The central test is whether those conditions can produce repeatable, heritable, and mechanistically traceable changes in host evolutionary trajectories.