view article
Figure 4
Performance of human crowdsourcing on two low-resolution phasing test cases. Plots showing the evolution of WRMS phase error during a game. In both the two-atom (a) and the cylindrin polypeptide (b) cases, the human-powered GA drove convergence towards solutions with much lower phase errors than random (pink). For comparison, the trend of the automatic GA using the WRMS phase error as a fitness function is also shown (green). The skewness for the whole population features an inverse correlation to the average WRMS phase error across the generations (grey).

Journal logoBIOLOGICAL
CRYSTALLOGRAPHY
ISSN: 1399-0047
Volume 70| Part 6| May 2014| Pages 1538-1548
doi:10.1107/S1399004714006427
Follow Acta Cryst. D
Sign up for e-alerts
E-alerts
Follow Acta Cryst. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS