Kinexus Bioinformatics is launching two new protein kinase microarray products based on data generated using its kinase substrate prediction algorithm.
The first, which Kinexus president and chief scientific officer Steven Pelech told ProteoMonitor will be released as an in-house service next week, is a peptide microarray for determining optimal peptide substrates for given protein kinases.
The second product, which the company aims to release in the next six months, is a peptide microarray designed to measure the activity of protein kinases in crude cell lysate.
Kinexus plans to offer that array in a chip-based format, marking the firm's entry into the microarray product business after having operated mainly as a service provider since its launch more than a decade ago.
"Over the last twelve years we've been a services company, primarily offering proteomics services, Pelech said. "And over the last two or three years we've been doing a lot more on the bioinformatics front. The next step is products."
The coming slate of offerings, which, according to Pelech, will also include the antibodies, protein kinases, and peptides used in the company's arrays, largely arose from the bioinformatic work the firm has focused on over the last several years.
In June, Kinexus announced that, in collaboration with researchers in the Mathematics of Information Technology and Complex Systems groups at the University of British Columbia and Simon Fraser University, it had developed an algorithm to predict which protein kinases are most likely to target a given phosphorylation site. The development has allowed the company to match kinases to the human proteome's presently known 92,000 human phosphorylation sites (PM 06/11/2010).
Since then, Kinexus and its research partners have developed another algorithm, this one for predicting putative phosphorylation sites across the human proteome's 22,000-plus proteins. With this algorithm they've predicted 650,000 phosphorylation sites, ranking for each the 50 protein kinases most likely to phosphorylate them.
The new peptide array products stem from Kinexus' efforts to experimentally verify these algorithms' predictions, which are available in the company's open-access PhosphoNET database along with additional information including evolutionary analyses of cross-species phospho-site conservation.
"Our prediction algorithm allows us to predict [substrate] specificities for protein kinases, but the question is, how do you know it's really working?" Pelech said. "We trained our algorithm using known [substrate] consensus sequences, but how do you establish that [it works] for kinases that you've never tested before?"
To solve that problem, Kinexus built the peptide arrays it plans to launch next week, generating synthetic peptides that correspond to the predicted optimal substrate sequence for each of the 500 kinases and putting them in a microarray.
They then took the protein kinases of interest and ran them on the array to see which peptides they phosphorylated, and at what levels.
Using that data, the company takes the peptides "where we get the best phosphorylation by the kinase," Pelech said. "When we align those sequences we can then derive a consensus sequence of what [substrate] that kinase is looking for – not by prediction, but by actual in vitro experimental data."
The company has run 25 protein kinases on the system and is in the process of running 100 more, after which it plans to publish its results. Thus far, Pelech said, the microarray-generated consensus sequences have matched the algorithmically predicted sequences more closely than the original consensus sequences generated through literature searches.
"The reason for that is that when you get the data from the scientific literature of what might be substrates, there's no quantification of that data – it's simply, 'Is this a substrate or not?'" he said. "Most of the substrates that have been reported in the literature are not necessarily good substrates, so your datasets are populated with data that actually dilutes out some of your important substrates."
Pelech said he envisions the assay will be useful for academic labs interested in studying particular protein kinases as well as pharmaceutical companies doing kinase-inhibitor research. The service will cost around $750 for four proteins, he said.
Looking forward roughly six months, he hopes to use the consensus sequences generated through this peptide array work to build the company's second new microarray offering — chips featuring optimal substrates for a range of protein kinases that can be used to simultaneously track the enzymatic activity of hundreds of kinases in crude cell lysates.
"The idea is that it's a broad screen, and if you see a dramatic difference between your disease state compared to your control tissue, then you have a peptide substrate probe from which you can go back and identify what are the most likely kinases being altered in your model system," Pelech said.
As part of the array work, Kinexus is also investigating what Pelech said is a novel strategy for generating the antibodies used in the arrays. He declined to elaborate on the technique but said that it was similarly guided by the company's bioinformatics programs and, if successful, could reduce the cost of making antibodies by five-fold.
"If you take a look at the cost of buying an antibody today, it's not that much different from even 20 years ago," he said. "You look at the cost of, say, having DNA sequenced, having DNA synthesized, having probes for these things – it's dropped dramatically – but for proteomics that hasn't happened. So we're trying to figure out if there's a way that we can improve the process for making antibodies."
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.