We argue that Jan Conrad's depiction of our preprint (http://arxiv.org/abs/1503.02320; 2015) as a case study in 'crying wolf' lacks accuracy and credibility (Nature 523, 27–28; 2015).

Based on public data from NASA's Fermi Large Area Telescope (LAT), we reported a γ-ray signal from the dwarf galaxy Reticulum II. Conrad characterizes our work as “the latest dark-matter discovery claim” and criticizes the “misuse” of public data at a time when an update from the Fermi collaboration “was imminent”.

Nowhere do we claim to have discovered dark matter. Rather, our paper is devoted to quantifying the probability that the observed signal is due to random fluctuation. Our closing paragraph says “it would be premature to conclude [the signal] has a dark matter origin”, then identifies future work necessary to establish such a discovery.

Our use of public data is concordant with the principles of 'reproducibility' Conrad invokes. Nevertheless, he compares our work unfavourably to a paper by the Fermi-LAT and Dark Energy Survey (DES) collaborations, who calculate a larger probability of background fluctuation (see preprint at http://arxiv.org/abs/1503.02632; 2015 and A. Drlica-Wagner et al. Astrophys. J. 809, L4; 2015). Conrad did not disclose that he was initially an author on their submitted paper. He states that the Fermi-LAT/DES result is based on “more comprehensive re-analysis of the same data”; however, theirs is a separate analysis of different data that were released 15 weeks after both papers appeared, preventing confirmation of their results in the interim. Meanwhile, only our result was reproducible (see, for example, D. Hooper and T. Linden, preprint at http://arxiv.org/abs/1503.06209; 2015).

Moreover, our findings are now published in the peer-reviewed literature (Phys. Rev. Lett. 115, 081101; 2015).