Comparing an Atomic Model or Structure to a Corresponding Cryo-EM Image at the Central Axis of a Helix
Description/Abstract/Artist Statement
Three-dimensional density maps of biological specimens from cryo-electron microscopy (cryo- EM) can be interpreted in the form of atomic models that are modeled into the density, or they can be compared to known atomic structures. When the central axis of a helix is detectable in a cryo-EM density map, it is possible to quantify the agreement between this central axis and a central axis calculated from the atomic model or structure. We propose a novel arc-length association method to compare the two axes reliably. This method was applied to 79 helices in simulated density maps and six case studies using cryo-EM maps at 6.4-7.7Å resolution. The arclength association method is then compared to three existing measures that evaluate the separation of two helical axes: a two-way distance between point sets, the length difference between two axes, and the individual amino acid detection accuracy. The results show that our proposed method sensitively distinguishes lateral and longitudinal discrepancies between the two axes, which makes the method particularly suitable for the systematic investigation of cryo- EM map-model pairs.
Presentation Type
Event
Location
Learning Commons @ Perry Library, Conference Room 1310
Start Date
2-18-2017 10:15 AM
End Date
2-18-2017 11:15 AM
Upload File
wf_no
Comparing an Atomic Model or Structure to a Corresponding Cryo-EM Image at the Central Axis of a Helix
Learning Commons @ Perry Library, Conference Room 1310
Three-dimensional density maps of biological specimens from cryo-electron microscopy (cryo- EM) can be interpreted in the form of atomic models that are modeled into the density, or they can be compared to known atomic structures. When the central axis of a helix is detectable in a cryo-EM density map, it is possible to quantify the agreement between this central axis and a central axis calculated from the atomic model or structure. We propose a novel arc-length association method to compare the two axes reliably. This method was applied to 79 helices in simulated density maps and six case studies using cryo-EM maps at 6.4-7.7Å resolution. The arclength association method is then compared to three existing measures that evaluate the separation of two helical axes: a two-way distance between point sets, the length difference between two axes, and the individual amino acid detection accuracy. The results show that our proposed method sensitively distinguishes lateral and longitudinal discrepancies between the two axes, which makes the method particularly suitable for the systematic investigation of cryo- EM map-model pairs.
Comments
Mentor: Drs. Jing He, Willy Wriggers, and Julio Kovacs