Durham, NC-based Expression Analysis, an Affymetrix-platform microarray-services provider, early this week announced that it has completed a “mock submission of microarray” data to the US Food and Drug Administration.
The company was retained to create the type of data package that companies might provide to the FDA as part of the drug approval process.
The submission, archived on “multiple CD-ROMs,” is based on data created for a toxicology study conducted by Schering-Plough and collected by Expression Analysis. The submission contains data on combined laboratory and informatics infrastructure, RNA standards, biological and processing factors, data-analysis methods, and data documentation and is published within an XML format.
The announcement is significant in that it comes at a time when the FDA is considering just how to accept and consider in its regulatory processes non-clinical genomic data, such as that obtained from microarray-based gene-expression analysis research.
The agency has created several gene-expression database projects. This one; a second, based on integrating Iconix Pharmaceuticals’ DrugMatrix toxicogenomics database in the FDA; and a third with Gene Logic.
Gathering the Data
The Schering and Expression Analysis collaboration takes toxicological data from a discontinued Schering drug project to create a “mock submission” that includes microarray, clinical-chemical, and phenotypic data. This mock submission is intended to provide a platform where the FDA can examine the format, content, and context for such data, Expression Analysis chief executive officer Steve McPhail told BioArray News’s sister publication, BioInform in a June 25 interview.
It also serves as a testbed for Expression Analysis to develop the work processes that will allow it, at some point, to provide this work on a services basis.
McPhail said the effort has been rewarding, but not simple.
“It’s not easy,” he said. “No one has done this before, so we are breaking new ground. This requires a significant amount of time and effort to put together this type of information in a format acceptable to the sponsor and the FDA.”
The project is slated for completion by October, and a final report is due in November. The project addresses a lengthy list of issues, including laboratory infrastructure, sample processing and array QC/QA issues, and experimental design and replication. Additionally, the project will include examination of data management issues such as format and file structures, linkage mechanisms between microarray data and other datasets, statistical analysis systems and software, and inference and modeling methods.
Expression Analysis used Affymetrix’s MAS 5.0 software to analyze the data, and its sister company, regulatory informatics firm Constella Group, handled the integration between microarray data and other clinical information.
Initially, the project followed CDER’s current guidance recommendations for regulatory submissions in electronic format, with the goal of identifying areas that need to be modified or redefined. This guidance stipulates that datasets be submitted as an SAS transport file of less than 25 MB per file, with data variable names of no more than eight characters, data elements defined in data-definition tables, and variable names and codes consistent across studies.
The submitted array data will include raw data files from image analysis. In addition, a summary report will be provided to describe normalization, data processing, and statistical analysis steps.
In conducting the project, Expression Analysis developed work processes and methods to stitch together disparate data, McPhail said.
“We had to get very innovative in working with different software languages, working with XML, and appropriate linkage mechanisms to link the microarray data with other data sets,” McPhail said.
The company had an eight-person taskforce assigned to the project and held weekly meetings to track progress. The next goals, McPhail said, include bias mitigation, and interpretation of the toxicology data set.
After the final report is submitted, the company will continue to work on compliance with the regulatory initiatives of 21 CFR Part 58 - GLP, and 21 CRF Part 11 - Electronic Records and Signatures, McPhail said.
The issue of microarray data submissions to the FDA is one that reflects the relative youth of the technology, and the hurdle that the microarray industry must overcome in moving from the laboratory workbench to the regulated clinical in vitro diagnostics market.
The agency convened a 14-member pharmacology and toxicology subcommittee for a one-day meeting in June to hear industry, academia, and drug sponsors speak about how regulators should consider microarray data in its drug approval process.
Currently, such data can be submitted with an Investigational New Drug application but rarely is, as there is no institutional framework of knowledge or policy guidance for how the agency can use this information. Thus far, the pharmaceutical industry has been averse to risking potentially negative regulatory consequences by sharing data in such an environment. Industry is concerned that without institutional guidelines for handling microarray data and recognizing its experimental nature, this data might cause the agency to order new clinical trials on a drug, or even stop a late-stage trial — both expensive propositions.
The subcommittee that met in June was to create recommendations to hand up to the advisory committee for pharmaceutical science, which then will craft the guidance documents that will go through a comment period and industry workshops.
The agency was to have submitted a draft guidance in August, but, according to industry sources, this guidance has been delayed. FDA representatives were unavailable for comment.