The Pervasive Effects of an Antibiotic on the Human Gut Microbiota, as Revealed by Deep 16S rRNA Sequencing

Your gut contains at least one hundred trillion microbial cells — that's ten to the fourteenth power — plus a quadrillion viruses. These organisms aren't passengers. They ferment carbohydrates that your body can't digest on its own, regulate immune cell development, and perform metabolic functions so central to health that some researchers have called the microbiota your "forgotten organ." And then you take a five-day course of antibiotics. That's the tension at the heart of a landmark study by Les Dethlefsen, Susan Huse, Mitchell Sogin, and David Relman. The question they set out to answer was simple: what actually happens to that microbial community when you take ciprofloxacin? But to answer it properly, they first needed a method capable of seeing the full community — and that turned out to be a problem. Before deep sequencing, microbial surveys relied on culturing bacteria in the lab or cloning individual genes and reading them one by one. Both approaches left most diversity invisible. Only about half of the most abundant gut bacteria matched sequences from cultured strains, and low-abundance organisms were almost entirely missed. Dethlefsen and colleagues used a different approach: pyrosequencing of the sixteenth S ribosomal RNA gene, a universal bacterial "barcode" present in all bacteria but variable enough to distinguish them. By targeting two hypervariable stretches of that gene — regions called V3 and V6 — and sequencing at massive scale, they could detect organisms that earlier methods simply couldn't find. The scale matters. The team generated more than seven thousand near full-length sixteenth S sequences by traditional cloning, then added over nine hundred thousand pyrosequencing reads — four hundred forty-one thousand eight hundred ninety-four from V6, and four hundred ninety thousand six hundred ninety-nine from V3, after quality filtering. These reads were grouped into reference-based operational taxonomic units, or refOTUs: clusters of sequences sharing the same nearest match in curated reference databases, used as practical proxies for species. The pooled dataset yielded three thousand three hundred sixteen V6 refOTUs and five thousand six hundred seventy-one V3 refOTUs across three individuals. Good's coverage — the probability that any randomly drawn sequence belongs to an already-observed group — reached ninety-nine point two percent for V6 and ninety-nine point five percent for V3, compared to just eighty-eight percent for the traditional clone libraries. That gap is the difference between a census and a sample. And the deeper census revealed a long tail of rare organisms: one thousand three hundred sixty-one of the V6 refOTUs and two thousand three hundred six of the V3 refOTUs were represented by a single tag. This is what researchers call the "rare biosphere" — and it turns out ciprofloxacin reaches into it. Ciprofloxacin influenced the abundance of roughly one-third of all detected bacterial taxa. That's not a minor side effect — that's a restructuring. Taxonomic richness, diversity, and evenness all fell during treatment. The Shannon diversity index — calculated by summing, for each taxon, its proportional abundance multiplied by the log of that proportion, then negating — dropped significantly. So did diversity measures weighted toward abundant taxa. The community didn't just lose rare members; its internal balance shifted. Specific losses were stark. A Clostridiales refOTU that ranked forty-first in abundance before treatment in one subject, averaging two hundred six tags per sample, was completely absent from every sample after ciprofloxacin began. A Bilophila-affiliated refOTU averaging sixty to ninety-one tags per sample in two subjects fell to fewer than two tags per sample after treatment. RefOTUs affiliated with Faecalibacterium declined in all three participants. Some taxa held steady or even increased — certain butyrate-producing groups like Roseburia and Anaerostipes — but the dominant story was collapse. Without pyrosequencing at this depth, most of this disruption would have been invisible. Here's where the story gets complicated. The biggest source of variation in these samples wasn't ciprofloxacin. It was the individual. Across all three subjects, interindividual differences in community composition outweighed the effect of the drug. Each person's gut microbiota had a distinctive fingerprint that persisted through treatment. At the fine taxonomic scale that deep sequencing enables, people look genuinely different from one another. And then the data throws a curveball. Two of the unrelated participants shared a surprisingly high degree of community similarity — which complicates any clean story about uniquely personal microbiomes. It suggests that while individuality is real, there may be stable ecological configurations that different people independently arrive at. The drug's effects were also person-specific in ways that defy simple summary. The Bilophila refOTU that collapsed in individuals A and C showed a different pattern in individual B: it was absent during treatment but rebounded to pretreatment levels afterward. Among the ninety most abundant V3 refOTUs, eight showed directly contrasting responses to ciprofloxacin across individuals. Even within a single bacterial genus, Bacteroides, eighteen of the nineteen most abundant refOTUs differed significantly between subjects — with a false discovery rate of just one point nine percent — even though the genus as a whole didn't differ significantly. The same drug, hitting the same genus, produced different outcomes at the species level depending on who's carrying it. That's the kind of resolution only available at this sequencing depth. Now, recovery. By four weeks after treatment ended, the taxonomic composition of the gut community in all three individuals closely resembled its pretreatment state. That's the headline. The community bounced back. Dethlefsen and colleagues interpret this as evidence for functional redundancy — the idea that different bacterial taxa can perform overlapping metabolic roles, so the ecosystem's core activities, fermentation of polysaccharides and production of short-chain fatty acids, can be maintained even when particular members are lost. The structure comes back because the functions have multiple performers. But the recovery was incomplete. Several taxa failed to return within six months of treatment. The Clostridiales refOTU that disappeared during treatment in two subjects hadn't recovered by the end of the study window. The Bilophila refOTU that collapsed in individuals A and C showed no recovery within that timeframe. The authors flag this with a specific example from the literature: Oxalobacter formigenes, a low-abundance bacterium that degrades oxalate, is associated with kidney stone risk when absent. It's a reminder that low abundance doesn't mean low importance. A taxon can be numerically minor and functionally critical — and a course of antibiotics might eliminate it without anyone noticing until years later. What determines whether recovery is complete? Dethlefsen and colleagues point to a mix of forces: competition within the community, dietary and host-derived environmental pressures, and recolonization from protected reservoirs like the mucosal layer of the gut. These are the factors that will need to be untangled to understand why some people's microbiota fully restore and others' don't. The study couldn't answer that question — it had three participants — but it precisely defined the question. The methodological legacy of this work is hard to overstate. By generating almost a million sequencing tags from time-series samples around an antibiotic course, Dethlefsen and colleagues demonstrated that the gut microbiota could be characterized with near-complete coverage, that change could be tracked longitudinally, and that rare taxa were both detectable and ecologically relevant. That framework — deep sixteenth S sequencing, time-series sampling, refOTU-based analysis — became the template for microbiome research going forward. The clinical frame shifts when you sit with these findings. Ciprofloxacin is one of the most prescribed antibiotics in the world. Most people who take a five-day course report normal intestinal function throughout. But at the microbial level, a third of the community is disrupted, rare organisms disappear, and some taxa don't return for at least six months — possibly longer, possibly permanently. The gut mostly recovers. But "mostly" is doing real work in that sentence. Who falls into the incomplete-recovery group? What does long-term perturbation of even a few taxa mean for immune function, metabolism, or pathogen resistance over years or decades? This study couldn't answer those questions. What it did was prove they were worth asking — with the precision to eventually answer them. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.