Breakout discoveries in science: what do they have in common?

We applied a series of computerised search algorithms to identify potential breakthrough discoveries in science. These citation-based 'early detection' algorithms extract their information from high-quality bibliographic databases. With early detection we focus on a paper's citation pattern during the first 24-36 month after publication. In this study we apply those algorithms to large sets of publications from 2007-2011 in the Web of Science Core Collection database, and generated a dataset consisting of publications that are considered high-impact papers. We call these papers 'breakouts'. We present general bibliographic characteristics in an attempt to classify communalities and develop a general typology of our identified 'breakout' publications in worldwide science. Our first results reveal a significant overrepresentation of various kinds of cross-sectorial collaborative research. The first results also reveal that 'breakout' publications are produced by on average larger teams of researchers and that this size depends on the type of collaboration.