I found a press release with slightly more info (in a human friendly language): <a href="https://cmns.umd.edu/news-events/features/3188" rel="nofollow">https://cmns.umd.edu/news-events/features/3188</a><p>I think this is the preprint: <a href="http://arxiv.org/abs/1506.08614" rel="nofollow">http://arxiv.org/abs/1506.08614</a><p>IIUC it's only a 2 sigma. It's not so rare that a 2 sigma just disappear with more data, because it was a statistical fluke. So don't get too attached to this result for now.