We are excited to add the first tool for automated model building into cryo-EM maps: Model Angelo.

In short, this tool uses advances in machine learning representation of protein structure with cryo-EM maps to leverage the combined information of chain tracing, amino acid geometry, and cryo-EM density simultaneously.

It works really well – in 10 minutes on COSMIC2 we were able to build a ~40 kDa model for KIFBP (a project in the Cianfrocco laboratory) for the published map EMD-24677 (!):

While subtle differences exist between the manually built model and Model Angelo, Model Angelo got the sequence registry and topology of the protein exactly the same as what we found after the painstaking work of model building.

Try it out! Model Angelo tool page

Preprint: ModelAngelo: Automated Model Building in Cryo-EM Maps. Jimali, Kimanius, Scheres. arXiv:2210.00006

We are excited to add a new tool to COSMIC2: Isonet Refine. Isonet is a deep learning-based data augmentation approach that tries to recover missing wedge information in cryo-electron tomography tomograms.

We are doubly excited to offer software that supports cryo-electron tomography (cryo-ET). We are keen to add new tools that allow cryo-ET to leverage computing at SDSC via COSMIC2.

To learn more about Isonet:

“Isotropic Reconstruction of Electron Tomograms with Deep Learning.” Yun-Tao Liu, Heng Zhang, Hui Wang, Chang-Lu Tao, Guo-Qiang Bi, Z. Hong Zhou. bioRxiv 2021.07.17.452128; doi: https://doi.org/10.1101/2021.07.17.452128. Link.

Are you struggling with setting up cryoEM IT infrastructure? Are your jobs dying for unknown reasons? Are you submitting a grant for computational equipment?

We can help if you have questions – as a cryoEM IT platform, COSMIC² is keen on helping others with the knowledge that we have developed over the years.

Head over to this page to submit a form and we’ll be in touch! 

We are excited to have finally launched the front end of COSMIC² at cosmic-cryoem.org! This site will serve as the launching pad for the user portal, as well provide a framework for information to help users get started using the high performance computing capabilities available to COSMIC².

We are also in the process of performing our first Relion job submission tests for hard coded jobs from the COSMIC² virtual machine hosted at SDSC. Once this is successful, we will be sure to report back : )