SBIR/STTR Award attributes
Experimentally obtained, structural models of macromolecules provide information that helps us to understand biochemical and chemical reactions and to design compounds that modulate these reactions. Therefore, methods that generate such structural models are highly significant to research and development efforts in industrial and governmental laboratories. Cryo-electron microscopy single particle reconstruction (cryo-EM SPR) is a powerful technique that can generate highly accurate structural models. In cryo-EM SPR, features of micrographs are analyzed and classified to enable alignment and signal averaging from thousands to millions of particle snapshots – each showing a molecule of interest in a single orientation. Success of a cryo-EM SPR experiment depends on the accuracy of alignment between single images of molecules, and for macromolecules having smaller masses, the amount of information in each snapshot is frequently insufficient for accurate alignment. Our project addresses this major barrier to cryo-EM SPR progress by developing and implementing new and powerful algorithms that use graphical processing units (GPUs) to recover and combine information present in cryo-EM images in the most optimal manner. In Phase I, we developed and implemented algorithms using GPUs that perform the early steps of cryo-EM SPR data analysis – classification and image alignment for macromolecules. Our solutions have definite superior aspects that we plan to leverage in Phase II, where we will incorporate our methods into an integrated software application. Our integrated application will include not only superior algorithmic and numerical approaches, but also elaborate management of a complex hierarchy of calculations, a graphical user interface providing fast, thorough, and informative feedback on project progress. This user-friendly package will be commercially distributed (Phase II and III). We expect that our novel algorithms and new approach to resource management will contribute to cryo-EM SPR’s fast expansion and will enable studies of macromolecules and their complexes that are currently inaccessible to the technique, either due to their size or their structural heterogeneity. These benefits are of high significance for the field, so the potential for successful commercialization is high.
