Gut Microbiome Competition: Trillions of Bacteria, One Shared Pantry

A shared, constantly refilled pantry

The human gut hosts an estimated hundreds to low thousands of bacterial species, all sharing the same physical space and, largely, the same incoming supply of digested food. Unlike a sealed laboratory flask, the gut is an open system: nutrients are continually replenished as the host eats, and bacteria are continually washed out through normal gut transit. This combination — shared, limited resources plus continuous replenishment and removal — is exactly the setup ecologists model with competition equations, and it is a useful lens for understanding why some gut communities are stable for years while others shift dramatically after a single course of antibiotics.

Generalized Lotka-Volterra competition

The classic Lotka-Volterra competition equations, originally developed for two competing species, generalise naturally to many species sharing resources. For each species i, the population growth rate depends on its own intrinsic growth rate, a self-limiting term (its own population crowding itself, bounded by a carrying capacity), and a sum of pairwise competition terms from every other species present:

dN_i/dt = r_i * N_i * (1 - (N_i + sum_j( a_ij * N_j )) / K_i)

N_i    population of species i
r_i    intrinsic growth rate of species i
K_i    carrying capacity of species i, set by the shared nutrient supply
a_ij   competition coefficient: how much one individual of species j
       suppresses species i's growth, relative to species i's own
       members (a_ii = 1 by definition)

The full set of a_ij values across every pair of species forms the interaction matrix, and it is this matrix — not the growth rates or carrying capacities alone — that determines the long-term fate of the community. A matrix where competition coefficients are large relative to self-limitation tends to produce competitive exclusion; a matrix where species limit themselves more than they limit each other tends to produce stable coexistence.

Competitive exclusion: the strict version

The competitive exclusion principle, formalised by Georgy Gause through laboratory experiments with competing paramecium species in the 1930s, states that two species relying on exactly the same single limiting resource cannot coexist stably: whichever species can survive on a lower resource concentration will, given enough time, drive the other to local extinction, even if the difference in their requirements is tiny. In the equations above, this corresponds to a case where the competition coefficients between two species are strong enough, relative to self-limitation, that the system has no stable interior equilibrium where both survive — trajectories converge to one species or the other winning outright, with the outcome determined by starting conditions and the exact coefficient values.

Why real gut microbiomes are more forgiving

Strict competitive exclusion would predict a gut dominated by a single winning species, which is not what is observed — real microbiomes routinely host hundreds of coexisting species for years. The resolution is that gut bacteria rarely compete for a single shared resource in the way the strictest version of the principle assumes. Niche differentiation spreads competition across many resource axes: different species specialise in digesting different classes of dietary fiber, different simple sugars, or different host-derived mucus glycans, which dilutes head-to-head competition on any one resource. Cross-feeding goes further, actively turning some competitive relationships into cooperative ones: one species' fermentation waste product — short-chain fatty acids, for instance — is frequently another species' substrate, so the presence of one species can directly benefit another rather than only compete with it. Continuous nutrient inflow and spatial structure along the length of the gut add further stabilising effects that a simple, fully-mixed, single-resource model does not capture.

Simulating coexistence versus collapse

A simulation built directly on the generalized Lotka-Volterra equations above, with a nutrient pool that continuously replenishes at a fixed rate and is drawn down by whichever species are present, reproduces both outcomes side by side depending purely on the interaction matrix chosen. Set competition coefficients close to 1 across the board — every species suppresses every other about as much as it suppresses itself — and the simulation reliably collapses toward one or two dominant species. Set most off-diagonal coefficients well below 1, representing strong niche differentiation, and the same equations settle into a stable multi-species equilibrium where every species persists at a nonzero population. Between these extremes, some coefficient combinations produce sustained oscillations rather than a fixed equilibrium at all — populations that rise and fall indefinitely, never quite settling, a reminder that even a small, deterministic system of competition equations can generate genuinely complex long-term dynamics.

Why this matters clinically

This framework is more than an academic exercise: it is the conceptual basis for understanding why a course of antibiotics can leave a gut microbiome vulnerable to colonisation by pathogens like Clostridioides difficile. Antibiotics disrupt the existing interaction matrix by knocking down populations that were previously suppressing a pathogen through resource competition, opening a temporary "niche" the pathogen can exploit before the original community re-establishes its competitive equilibrium — precisely the dynamic these equations are built to describe.

Frequently asked questions

Try it live

Everything above runs in your browser — open Gut Microbiome Competition and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.