NEW YORK (GenomeWeb) – Researchers at Finland's Turku Centre for Biotechnology have received a five-year, €1.5 million ($1.7 million) grant from the European Research Council to fund development of bioinformatics tools for longitudinal proteomics.
Using the award the researchers plan to improve upon existing tools for analysis of longitudinal proteomics data, studying the issue in the context of type I diabetes, Laura Elo, research director in bioinformatics at the TCB and leader of the project, told GenomeWeb.
Elo noted that she has been involved in proteomics research for some time as well as longitudinal analysis of other data sources and that recent improvements in proteomic technologies have made more feasible the longitudinal collection of deep proteome profiles.
"Technologies have been limited in terms of being able to produce large-scale longitudinal proteomics data, but now [such data] is increasingly becoming available." she said. However, she added, the informatics tools required to analyze such data are not yet fully in place.
Longitudinal analysis is an area of interest to biomarker and diagnostics research generally and proteomics specifically, with one of the main potential advantages being the possibility of accounting for natural variation within a population by using individuals as their own baselines.
The notion is not a new one outside proteomics, Elo said, noting her work looking at longitudinal data from other omics disciplines as well as from more conventional medical records. And, within proteomics, a number of researchers and companies are looking to incorporate longitudinal analyses.
For instance, earlier this year, researchers at Arizona State University's Biodesign Institute published a study in Molecular & Cellular Proteomics that showed that immunosignatures of healthy human subjects generated using their peptide microarray-based technique remained constant over time, a finding that, the researchers noted, suggested longitudinal testing of individuals using the method could prove effective for early detection of disease.
Similarly, clinical proteomics firm SISCAPA Assay Technologies has been exploring the use of its technology in combination with dried blood spot sampling for longitudinal monitoring of individuals. As Leigh Anderson, SAT's founder and CEO, noted at the Mass Spectrometry Applications to the Clinical Laboratory annual meeting in 2014, biological variability within populations can make it difficult to determine appropriate biomarker cut-points for individuals, an issue that longitudinal testing could address.
More recently, Anderson and SAT have used longitudinal protein measurements to track the progress of Brazilian athletes at last year's Pan American games in Toronto. The company is also working on developing the approach to track the progression of infections in individuals.
These efforts differ from the ambitions of Elo and her colleagues in that they involve fairly targeted assays focused on a relatively small number of analytes. The TCB researchers are looking to apply longitudinal analyses to shotgun-style proteomic datasets, which consist of thousands of proteins, a context in which Elo said informatics tools have not been fully worked out and tested.
The goal, she said, is to combine the temporal information provided by longitudinal approaches with the sort of complex pathway data provided by large-scale omics studies. Tools exist for longitudinal analyses, but, Elo said she believed they were "not really tuned to work in the proteomics context. Or, at least, people do not know yet whether they work [in this context] or not."
"So borrowing from conventional statistics on the one hand, and from omics studies which have used these pathway [analysis] types of methods," she said. "We aim to develop tools to generate reliable data in this area, and in terms of putting these two aspects together, I think there is still quite a lot to do."
As the research progresses, there is a possibility that Elo and her colleagues will try to develop more targeted protein panels for studying type I diabetes, "but at the beginning I do not want to limit what we are measuring," she said.
The number of patients they follow and the number of time points they collect data on will be determined by their ability to access samples, but Elo said that she would hope to collect data from at least five time points and "preferably much more." She added that the researchers have access to patients from an ongoing Finnish study of type I diabetes launched in 1994 that has sampled participants every three months. "So it is quite dense," she said. "And I think we will learn during the project how dense we should go."
Type I diabetes is an ongoing area of study for Elo, which is why the researchers are using the disease as their model in developing longitudinal proteomics data analysis tools. The disease has very high incidence in Finland, she said, noting that it has a strong genetic component and that the studies are following children deemed to be at high genetic risk of developing the disease.
In addition to genetic indicators, certain autoantibodies also commonly appear in the blood of patients prior to diagnosis, Elo said. "But it is very difficult to know when the actual disease comes. So it is important to [be able to detect] it as early as possible."
The grant started in June and Elo said that she is currently recruiting researchers to work on the study. Elo also received in May a €500,000 grant from the Academy of Finland supporting work on the development of bioinformatics tools for individualized prediction of disease risk and response to treatment.
"I really hope that having done the project, we have a versatile and easy-to-use tool to analyze these longitudinal proteomics datasets," she said of the ERC-funded work. "And not only in the context of diabetes but in other diseases, as well."