Computers to extract huge amounts of relevant data

“We want to continue to grow internationally,” says Dr Tim Hiddemann, co-founder and CEO of antibodies-online GmbH in Aachen, Germany. Together with Dr Andreas Kessel, he started the company as a spin-off of the RWTH Aachen in 2006. By now, they offer 860,000 antibodies and related products, independent from vendor, on their online-marketplace antibodies-online.com. 2,000 research institutes in 53 countries, among them Harvard University and the MIT in Cambridge, USA, are part of their customer base.

French, Chinese and English platforms were launched at the end of 2007 and in mid-2008. To meet to the rapid growth and expansion of its customer base, a US subsidiary was founded in Atlanta in 2010 and an office was opened in Shanghai in 2013.

Antibodies are essential components of the immune system. As is commonly understood, they develop after a vaccination against specific diseases. Antibodies are therefore an important component for research and medicine, as they help identify proteins or give evidence of their absence. All pharmaceutical companies and countless scientists all over the world need them to develop medications against diseases such as cancer or rheumatism. There are at least one million antibodies on the market – which are now, thanks to antibodies-online.com, also easily available in the UK. But each scientist only needs specific ones which exactly match his or her particular experiment. The antibodies-online GmbH finds those needles in the haystack for the researcher and delivers them to the laboratory.

There are countless numbers of antibody varieties in thousands of catalogues and websites by suppliers across the globe. For that reason, one third of the employees at antibodies-online are software developers, constantly developing new programmes to make such varieties transparent and comparable. The ultimate goal is to automatically identify the most exact products required by the researchers as soon as possible.

How is relevant information filtered from language? For a human proof-reader this often happens intuitionally without much effort. However, the extraction of information constitutes a bigger challenge for information processed digitally or via computer. In collaboration with the FH Aachen (University of Applied Sciences) and the energy firm Ene’t GmbH (Ltd) from Hückelhoven, antibodies-online is developing new platform technology which is meant to help view relevant content from diversely structured sources and to transform these into the required data-model of any platform.

The joint project called ETL Quadrat is funded by the German Federal Ministry of Education and Research with more than €600,000. The aim of this project is the exploration of a new platform technology which helps coherently format extreme, unstructured amounts of data quickly and easily.

Antibodies-online employee, Marcel Ahne initiated this project using the so-called ETL process. This is carried out by accumulating data from several differently structured sources in one target database. Appling this model which originated from the field of finance, also helps to process information about antibodies. The research project with the FH Aachen and the enterprise from Hückelhoven now focuses on developing a new way to interpret language. Huge unstructured amounts of data could ultimately be scanned faster with regard to their content and could be edited into structured, automated information for online customers: in five years, the ETL-project, which is supported in the context of the funding initiative KMU-innovativ, shall be finished.

For companies that receive a huge amount of data which need to be edited from different suppliers every day, the completion of this project should prove particularly interesting.

The editorial unit