According to the passage, reverse causation is the default expectation in microbiome research because

illness alters diet, medication and gut transit, each of which reshapes the microbial community.

sequencing cannot reliably distinguish between closely related microbial species.

healthy control subjects are difficult to recruit in sufficient numbers.

microbial communities vary more between individuals than within a single individual over time.

The discussion of the gnotobiotic mouse in the second paragraph serves primarily to

propose a design capable of replacing correlational studies entirely.

establish that murine and human immune systems are fundamentally similar.

credit the field with having resolved the problem raised in the first paragraph.

show that the design meant to establish causation carries assumptions that narrow what it can prove.

It can be inferred that a phenotype which travels with a transplanted community demonstrates

that the donor's condition was caused by the donor's microbiome.

that some microbial property suffices to produce that phenotype in an atypical host, which is less than is generally claimed.

that human and murine gut communities are functionally equivalent.

that the phenotype would transfer between any two hosts.

The author's attitude toward the recent work described in the final paragraph is best characterized as

skeptical that it represents any improvement on transfer experiments.

troubled by its apparent abandonment of causal ambition.

indifferent, presenting it as one option among several.

approving, on the ground that it trades effect size for interpretability.

Which of the following, if true, would most weaken the author's criticism of transfer experiments?

Human-derived communities colonize germ-free mice with a fidelity approaching that of the donor gut, and immune development in the recipients proceeds normally.

Germ-free mice are considerably more expensive to maintain than conventionally raised animals.

Several human conditions associated with the microbiome have no murine analogue at all.

Interventions using defined consortia in humans have so far produced only small effects.

←

The Microbiome and the Problem of Causal Inference

PRACTICE

100:00+4 −1MCQ Single Answer

Reading Passage

Cheap sequencing transformed the study of the human gut. Within a decade of the technology becoming routine, the literature held hundreds of reported associations between the composition of an individual's microbial community and conditions ranging from obesity and depression to autoimmune disease. The volume of these findings was widely read as evidence of the microbiome's causal reach. It is better read as evidence of how easy the associations are to obtain. Composition differs between the sick and the healthy for many reasons that have nothing to do with the microbiome causing anything: illness alters what a person eats, what medication they take, and how quickly material moves through the gut, and each of those reshapes the community that lives there. In this domain reverse causation is not an exotic possibility to be acknowledged and set aside. It is the expectation any positive finding must be shown to defeat. The field's answer was the gnotobiotic mouse. An animal raised free of all microbes is colonized with a community transferred from a human donor; if the donor's phenotype travels with the transplant, the community is taken to have caused it. The logic is genuinely powerful, and it has produced results no correlational study could. But the design carries assumptions that are seldom stated alongside its conclusions. A germ-free animal's immune system develops abnormally, so the host receiving the transplant is not a typical host. The recipient eats a mouse diet, not the donor's. And only a fraction of human taxa colonize a murine gut at all, so what is transferred is not the donor's community but a selection from it. A phenotype that travels therefore demonstrates that some microbial property is sufficient to produce it in an atypical animal — a narrower claim than the one that is usually reported. The most informative recent work neither abandons transfer experiments nor treats them as settled proof. It pairs them with interventions in humans in which a single variable can actually be moved: a defined consortium of strains, a specific dietary fibre, administered against a control. The effects recovered this way are smaller than the mouse literature would lead one to expect, and the claims that survive are correspondingly modest. That is the point. The difficulty with microbiome research has never been that it is uninformative; it is that its most cited claims rest on a design whose limits are known to specialists and routinely elided in summary. Progress in this field will look less like a widening of scope than like a disciplined shrinking of ambition.

The primary purpose of the passage is to