Folding Domain Prediction Program
A large fraction of full-length proteins consists of multiple fold domains. To predict the boundaries of these fold domains for the whole proteome of an organism, we have developed a computation method, which provides all the predicted fold domains of the organism, and their homologues in other organisms based on their similarity to known domains in the SCOP and Pfam database.
Robotic Automation of High Throughput Cloning
To clone a large number of genes from genomic DNA or cDNA libraries, and to subclone each gene into multiple expression vectors, we have sub-divided the entire process from PCR to mini-expression screening into nine steps. All steps are robotized using a 96 well format.
Optimum Solubility Screen
To increase the success rate for obtaining crystals for structural studies, purified proteins are screened in a wide variety of conditions (such as pH, ionic strength, and additives) to find the solution conditions in which the proteins are soluble and homogeneous.
On-Column Refolding
We have developed an on-column screening method for finding refolding conditions for insoluble, sparingly soluble, or “poorly behaving” proteins as well as membrane proteins.
Improved Auto indexing of Diffraction Data
We have found that current auto-indexing programs often encounter problems, which render them impractical for use in automated data processing applications. Therefore, we have developed improved methods for indexing diffraction patterns from macromolecular crystals available in the Lawrence Berkeley Lab Indexing Toolbox (LABELIT).
Automatic Crystal Mounting and Crystal Screening
Robotic and computational processes have been developed at the Advanced Light Source, Lawrence Berkeley National Laboratory, to automatically screen many crystals either for their quality assessment or for contiguous collection of multiple data sets.
X-ray data repository
We have developed a web-based system to provide some aspects of data management when working on many different protein targets. The system makes use of Python, MySQL and Apache to provide a resource for storing integrated diffraction data and running automated programs such as HySS, SOLVE, and RESOLVE.
Mapping Protein Fold Space
To obtain a global view of protein fold space we have developed a computational method to map all known protein folds in a 3-D space, where similar folds are placed close to each other and four known fold classes are well segregated in space.
