Hi Cedric,

Is there a way with phenix, in a badly defined electron density area, to generate a set of different alternative conformations for a small protein fragment ?

you can run many dynamics trials each one with different random seed and this will generate many different models, for example like this: http://phenix-online.org/phenix_data/afonine/fem/fig_24_25_26_rsr_3I9Q/fig_2... Command: phenix.dynamics model.pdb random_seed=3456879 temperature=600 number_of_steps=500 stop_at_diff=0.5 Here: temperature and number_of_steps will control variability of models, stop_at_diff will terminate the job if RMSD between starting model and current one is greater then defined value. Pavel