IHM Meeting Summary May 18 2017

IHM Meeting Summary May 18, 2017

Attended by BW, TG, JW and BV

1. BV will add documentation in the dictionary regarding conventions to be followed while referencing 2DEM class average images and the corresponding transformation matrices (discussed in issue 30). This documentation will address the following points.

   • The lower left image corner (image pixel index 0,0) should be at x,y,z = (0,0,0).
   • The image should be in the XY plane.
   • The 2D image is scaled by the pixel_size_width and pixel_size_height from the ihm_2dem_class_average_restraint table.
   • The tranformation is applied after scaling and hence the translation vector should account for the scaling.
   • Clarify whether the model should be moved to fit the image or if the image should be moved to fit the model i.e., should the matrix be inversed?
   • For 2D images, are there specifications for Z translations i.e., how far the image should be from the model while projecting? Perhaps, not.

2. It is preferable for large sphere model ensembles to be in binary format. This facilitates faster loading. MMTF is the preferable binary format but it does not currently support coarse-grained I/H models. So, as a temporary solution, a binary dump of co-ordinates in dcd format can be used. The topology comes from the ihm_sphere_obj_site and atom_site tables. Refer to the corresponding IMP issue.