Dear Consortium Members and Affiliates,
SBGrid is in sync with the seasons; that's my explanation anyway. With spring in full swing, we have to step in time, and news abounds as SBGrid blossoms and grows. Read on to hear the latest on the SBGrid Data Bank, HADDOCK'S models dataset, a question about usage of your 32-bit software branch, a GROMACS webinar to watch and a SITUS presentation to attend, a scheduled start for our Schrodinger working group, a story on developer Geoff Barton, 3 new members to welcome, 3 member/publication highlights, and, finally, a software push that includes 8 updates and 2 new software titles.
The SBGrid Data Bank is now accepting historical and ongoing x-ray diffraction datasets from SBGrid member PIs. The response from the community has been overwhelming. In the last few weeks we have collected 90 datasets from 50 different groups. We received multiple emails from SBGrid labs reporting that old data cannot be located or that the media used for storage is no longer accessible. The SBGrid Data Bank will provide persistent storage for all members, so please do not delay and upload your datasets for published structures before they vanish.
We are also now accepting datasets for upcoming publication, for which the SBGrid Data Bank offers HOLD functionality. A few groups have already piloted this system and once datasets are deposited members can cite them in scientific publications. To deposit your datasets please visit http://data.sbgrid.org and read the Guidelines and FAQ sections for additional information about the deposition process.
The SBGrid Data Bank can also preserve data types that aren't supported in existing repositories. SBGrid member Alexandre Bonvin from the University of Utrecht recently uploaded a HADDOCK models dataset containing decoys for 55 new complexes in the protein-protein docking benchmark version 5. This updated benchmark was developed in a collaboration between the Weng laboratory (UMass Medical School), the Fernandez-Recio group (Barcelona Supercomputing Center), the Bates laboratory (Cancer Research UK), and the Bonvin laboratory (Utrecht University). The HADDOCK decoys were obtained using bioinformatics predictions, with their CPORT server, or defined CDR loops, in the case of antibody-antigen complexes, to drive the docking. Each set contains, respectively, 10000/400/400 PDB models corresponding to the three stages of the HADDOCK protocol (rigid body docking, semi-flexible refinement, final refinement in water), together with their associated CAPRI quality metrics (interface RMSD, ligand RMSD, and fraction of native contacts). This decoy set, which contains both near-native models and wrong models of high quality, should be valuable to groups developing new energetics/scoring functions to discriminate near-native complexes from wrong predictions.
Are you still using your 32-bit linux branch ? No, you're probably not. We can remove that branch from your installation and save you some valuable disk space, just let us know. Don't worry, we'll keep it handy and can add it back upon request if you need it.
In webinar news, we had a good turnout for our GROMACS webinar last week and have just posted the video to our YouTube Channel. Thanks to Erik Lindahl for his presentation. June will bring a visit from Willy Wriggers and a primer on SITUS, an application used for density docking purposes. Please be sure to join us on June 9th at 12pm EDT.
We'll also be kicking off a new Schrodinger Data Club/Training working group, for those of you that want to learn more about this software suite, which is available to all non-profit groups in North America and Australia. The first in this series of online meetings is scheduled for Tuesday, July 7th at 12pm. The session will start off with a 30-minute presentation by Vice President of Application Science, Woody Sherman, followed by a 30-minute Q&A session, where users can share workflows and seek advice. To participate in this working group, please complete our Google Form.
For our May webtale we are giving a turn to software developers with a look at Jalview developer Geoff Barton, Head of the Division of Computational Biology at the University of Dundee. Barton gave us peek at Jalview features that are on the horizon and shared one key lesson learned about successful scientific software: “Make it easy to use,” he says. “And give it away for free.”
Many of you have asked about SBGrid service renewals. We'll be sending invoices for the July renewal on July 1st, so please be sure to let us know if there are any new PIs at your site that wish to join, or if you need a formal quote in advance to satisfy your purchasing department.
May brought us three new members. Katherine Henzler-Wildman became our ninth member at the University of Wisconsin-Madison, Egin Ozkan our 10th at the Unviersity of Chicago, and Tim Clausen joins as our first representative from the Research Institute of Molecular Pathology (IMP) in Vienna. Welcome to our new members!
For the month of May our software push includes updates to ARP/wARP, CCP4, CTFFIND4, EPMR, MAFFT, Phenix, Rosetta, and Schrodinger, along with two new software programs: DIALS and SHIFTX2. Please read below for more details.
Member Publications
SBGrid member labs published over 50 manuscripts during the last month, a full listing of which can be found on the Member Publications page of the SBGrid website. Here are some notable highlights:
- In his PLOS ONE paper, Doug Daniels from the Broad Institute adds to our understanding of MCL-1 by using an MCL-maltose-binding protein fusion platform to obtain the first high-resolution apo structure of MCL1.
- Kevin Corbett's work on TRIP13 revealed that this little protein gets help from adapter protein p31(comet) to convert the MAD2 protein from its signaling-active state to its inactive state. Read more at eLife.
- We have new insight into the nuclear pore complex, from Thomas Schwartz's MIT laboratory where he determined a 4.1-Å crystal structure of the 'hub" of the Y complex. More details on the Nature Structural and Molecular Biology site.
If you have news from your lab that you would like to share with the community, please let us know by emailing Michelle Ottaviano or communicating with SBGrid on Twitter (@SBGrid).
Software Changes
ARP/wARP version 7.5 was updated to include patch2. Please note, ARP/wARP is available to all non-profit labs, but the developer requires each user to register. Please see our website for more details if you would like this application included in your installation.
CCP4 version 6.5.010 stamped out bugs in several components, fixed an issue with degenerate PDB files, allows real-space reindex operators in pointless, and made ABSORPTION imax work in aimless.
Version 4.0.15 of CTFFIND4 fixes a bug that was causing multiple instances to write to the same file when running CTFFIND and relion together.
EPMR was updated to version 15.04. No release information is available on the website.
DIALS version dev20150513, a new software package for the analysis of crystallographic diffraction images, was just added to the SBGrid collection. This version of DIALS includes xia2 and the developers have indicatd that they will now publish xia2 updates through DIALS, with xia2 acting as the user friendly interface to DIALS for synchrotron applications.
In MAFFT version 7.220 the developers fixed a problem that shows itself when almost identical sequences are subjected to the iterative refinement options.
Phenix's dev-2037 version fixes a bug in NCS constrained refinement and DNA/RNA specific restraints and adds an option to specify rigid groups for rigid body refinement in phenix.real_space_refine.
Rosetta version 2015.19 is now the default.
Schrodinger's 2015-2 release is now available. You'll find that Maestro is much speedier, with improved displays and ease of movement. There is added support for running simulations on Nvidia Tesla K-80 GPUs, an option to dock ligands directly from a Phase database including subset support, and a number of other usability improvements and performance enhancements across all software.
New to SBGrid is SHIFTX2, version 1.08, a program for predicting the backbone and side chain 1H, 13C and 15N chemical shifts for proteins using their structural (PDB) coordinates.
Please don't forget to cite our eLife publication (eLife 2013;2:e01456) for all projects completed with SBGrid compiled software.
Please note that not all software applications are available to every SBGrid member type. If you see an application that you would like to use, but is not included in your software tree, please contact us to find out what options are available for access.