Tatjana Abaffy ^* James R. Bain Michael J. Muehlbauer Ivan Spasojevic Shweta Lodha Elisa Bruguera Sara K. O'neal So Young Kim Hiro Matsunami

• Duke University, Durham, North Carolina, United States

In the published article, there was an error in the legend for Figure 1 as published. Mistake was made in data normalization calculations and in the legend description. "Figure 1. Discovery of novel endogenous metabolite-ligands for OR51E2. A. Study design. B. Homology model of OR51E2 with 19-hydroxyandrost-4-ene-3,17-dione (19-OH AD) docked into the receptor pocket. C. Concentration -response curves for 19-OH AD and AFMK (red) and their structures. D. Concentration -response curve for antagonist 13-cis retinoic acid (13-cis RA). N=3, mean ± SEM."The corrected legend appears below. The authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated. In the published article, there was an error in the Figure 1B as published. A mistake was made in presenting a docking site in the homology model, Fig. 1B. Since we did not analyze docking of 19-OH AD on the receptor, we cannot be sure about the exact binding site, thus we conclude that it is inappropriate to present it in the Fig. 1B.A mistake was made in the presentation of concentration-response curves in the Fig. 1C. A mistake resulted from the incorrect calculation of data normalization. The concentrationresponse curves were presented in a wrong way, where 0 response with no chemical was presented as 0, instead of 1. This error was also corrected in the Suppl. Fig. 2,3 and 4 as discussed below.In the published article, there was an error in the Figure 1D as published. A mistake was made to claim that 13-cis RA is antagonist, while only concentration-response curve for its action as an inverse agonist was presented. The corrected Figure 1D Concentration -response curve for 13-cis retinoic acid (13-cis RA) in the absence (left panel) and in the presence of 100 nM 19-OH AD. Experiments are performed in M10d medium (left panel) and in the CD293 medium (right panel). Data are normalized to the response with no 13cis RA, N=3, mean ± SEM. 13-cis RA acting as an inverse agonist (left) and an antagonist (right).The authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated.In the published article, there was an error in Table 1 The authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated. In the published article, there was an error in Supplementary Figures 2, 3 Concentration -response curves for agonists (red) identified from the mfScore list (Table S1) and from the Score list (Table S2). Responses are normalized to a no-metabolite control. A pCI vector-only control is also presented for each metabolite (black dots). N=3, mean ± SEM. N = 3, mean ± SEM.