Computational discovery and refinement of targets

Our proprietary algorithms can be used to discovery new targets or rationally design conformational epitopes. Using PsiPhi™ we can identify conserved curvature between proteins and leverage that to engineer an epitope onto a scaffold with different background. This can be especially helpful with cryptic sites that only adopt a certain structure for a short time. If that structure is of commercial or scientific interest, grafting it into a different protein can be a method to “freeze” the conformation so that it can be used as an immunogen or in catalysis.

Rotational Angles Psi and Phi.

PsiPhi™ defines distances based on the Ramachandran map and screens a database for matches. In the figure below, a screen performed at Protein AI demonstrates how the scaffolding is matched. The template (left) is used to screen for closest match and identifies a target (right) that has very similar geometry to the original template.

A database screen with PsiPhi™.

Our second platform is based on standard bioinformatic searches combined with neural nets to search a genome to identify antigenic epitopes, including repeat sequences and T cell epitopes. Here we screen and rank targets, we identify full length sequences and predicted cellular location wherever possible and confirm the specificity of these sequences by evaluating against the genomes of other organisms. Following target ranking we can assist in the design, expression and purification of recombinant target proteins.

Target antigens made at Protein AI can be used to diagnosis patients.

Proteomic Compression

This technology is a method for the compression of large sets of biological sequence data into a minimum number of degenerate pools enabling the rapid processing of large data sets which cannot be efficiently processed using existing approaches. In the case of oligonucleotides this approach allows the creation of degenerate primers for PCR and hybridization experiments for much larger data sets than existing methodologies. In the case of polypeptides, the FlexGrePPS ™ allows for design of degenerate peptide mixtures to be synthesized for antigenic, diagnostic and functional screening. The identified pools are then synthesized and the resulting synthesized pools are then utilized to screen for targets. These targets are useful in the development of diagnostic and therapeutic products, vaccines, antibodies, as well as for use in general research, or as development tools, or biomarkers, or in diagnostic or therapeutic formulations, methods or kits. In addition, we hold intellectual property on a method of directed synthesis of degenerate pools of molecules to isolate target sequences. Producing and screening entire pathogenic proteomes in the laboratory is a time-consuming and prohibitively expensive process. We can reduce sets of protein sequences to degenerate peptide pools that lend themselves to rapid and economical screening and thereby promise to enable the experimental processing of entire proteomes.

FlexGrePPS™ performance as a function of sequence length and number of sequences.

Development services – from target identification to lead validation

Once a target is identified our wet labs can design the expression constructs and produce them either through PCR or synthetically at a gene synthesis company. The coding sequences are then inserted into a commercially available vector for expression in E.coli or other host as needed. We grow cells, harvest, and isolate proteins after lysis. Target purities are around 95%. The final product is assessed by standard biochemical means and can be validated by enzymatic or immunological assay. The figure below shows the process on a lead identified in house.

Target identification to lead validation. A target is identified and characterized in silico. Our wet lab produces the protein and performs immune readouts to validate the target.

Under development: Cloud computing and research services GeneCloud.com

The computational analysis of biological sequences is undergoing a revival that is fueled by the rapidly growing amount of next generation sequence (NGS) data. Analyzing these tremendous volumes of data is a time-consuming process that not only overwhelms traditional computational hardware infrastructures, but also renders inadequate most existing sequence analysis tools, which are based on single processor algorithms. Rapid and economical analysis of the sequence data, however, is crucial for a speedy translation of raw sequencing results into biological discoveries leading to much needed biomedical advances.

We see the strong need for a user-friendly tool that would harness the power of cloud computing to enable a broad spectrum of biomedical researchers to make swift progress in the stages of their research following the acquisition of raw sequence data. At the same time, we see think that the market adoption of our tool, and therefore the impact of the proposed project, will be significantly higher if the tool is able to easily communicate with existing and widely used resources in a flexible manner.

An example of implementation of such an interface is shown below. The user is directly interacting with the intuitive web interface of our GeneCloud portal. The Internet provides the connection with the CloudBurst read-mapping algorithm that is implemented in a cloud computing setting. Interoperability with R/Bioconductor, a widely adopted package, can be maintained via rJava/JRI/rserve.

The layout of GeneCloud. While the user is interacting with the web interface of GeneCloud, interoperability with R/Bioconductor can be maintained via rJava/JRI/Rserve, and the Internet provides the connection with CloudBurst that is implemented in a cloud computing setting.

Services Overview

Using our proprietary algorithms and our expertise in wet lab development and validation we offer a number of services to companies and researchers. These include:

• in silico target design and discovery
• Ligand screens and designer antigen
• in vivo studies, recombinant protein production and read-outs
• Partnerships with commercial entities for distribution of our leads

Our lab studies are performed as fee-for-service agreements and commercial partnerships operate on license-based model. Email us for details.