Drawing on two ethnographic examples from a laboratory study of a group conducting environmental epigenetics research on suicide risk, we examine the ways in which researchers go about making credible claims in the face of a range of profound uncertainties. We first explore how a range of factors led to what is now accepted as a fact or discovery being explained away, several years ago, as a combination of technical error and known background noise. To set this first example within a broader context, we turn to a debate that erupted in the lab during a journal club meeting about claims-making and publishing in science. Through these examples, we aim to demonstrate the complex terrain in which scientists manage epistemic uncertainties and produce credibility in environmental epigenetics research. Our goal is to trace uncertainties from the unknowable reasons and internal states that lead people to respond to data in a particular way, through to the technical difficulties in identifying data from noise, through to issues of ethical self-making, as students learn to become certain types of scientists. We argue that the ways in which uncertainty takes on meaning, has effects and is managed also has much to do with its situatedness.