Meet our Scientists
Wolfgang Thielemann: Finding the Untraceable
We have built hundreds of thesauri. Some of them contain more than 100,000 phrases. This way, we train our software to extract the essence of unstructured texts and to identify relevant relationships.
Patterns and structures have always fascinated me. In my spare time, I take photos of such configurations, whether I find them in rock formations, architecture or landscapes. I also work with patterns and structures professionally. I’m an expert for text mining and analysis in pharmaceutical research. My colleagues and I develop algorithms and workflows to systematically scan for and evaluate valuable information in scientific databases. We hunt for patterns in a great, messy pile of digital files.
Sifting Through Discoveries
Text mining helps in pharmaceutical research: These days, worldwide journals publish tons of scientific papers. It’s impossible for researchers to read them all. However, a machine has no problem in working through more than 10,000 publications per day. In addition, a machine doesn’t miss rare or tiny details – and it may help to find scientific trends. Today, such automated systems are essential. The development of drugs takes more than a decade, and during this time, scientists make a number of decisions. If they don’t have all relevant information, it may cost the company a lot of money.
To prevent this from happening, we start with questions: Do other drug companies have the same goal as we do? Do they have a lead compound already? How risky is it for us to invest in our own candidate? We can find initial answers to these questions in the public domain, such as in patent applications. But it’s not as easy as it seems. The information is often buried in extensive descriptions and claims. This situation makes it challenging to find the right clues about the progress of competitors’ compounds since standard keyword searches can easily miss the most relevant information. Our algorithms are constructed to deal with this situation and address these challenges. Furthermore, text mining provides solutions to specific experimental problems. We find publications linked precisely to the same problems from all around the world.
Finding New Passions
I always wanted to work in a lab. It was my dream since I was eleven-years-old, when I received a chemistry set. This set changed everything: I spent a lot of my free time doing experiments and invested all of my allowance on more materials and equipment. I also started recreating experiments that I found in my father’s old school books. Eventually I was quite ahead of my school curriculum, so in free periods my teachers invited me to give lessons to students in higher grades. If someone had told me then that I would end up searching and analyzing patent information, I would have doubted their sanity. I wanted to work in a lab.
After my chemistry studies in my hometown of Muenster, Germany, I had the most exciting year working as a postdoctoral researcher in Berkeley, California. Afterwards, I could have found work in the US, but I wanted to pursue my career in Germany. When I arrived at Bayer in Wuppertal in 1999, I was exactly what I had always wanted to be: a lab head in pharmaceutical research.
However, three years later, I moved to my current field, which was in its early stages and offered an interesting challenge. We had to program everything from scratch. Ever since, we’ve been dynamically developing this discipline. As well, I work with all kinds of people, from lab heads to chairmen. This makes my work so fascinating. I wouldn’t change a thing.
CV: Wolfgang Thielemann
1968 Born in Muenster, Germany
1989-1994 Chemistry studies, University of Muenster
1994-1997 Dissertation in organic chemistry, University of Muenster
1998 Postdoctoral fellowship, University of California, Berkeley
1999 Lab head Medical Chemistry, Cardiovascular Research, Bayer HealthCare, Wuppertal, Germany
2002 Head of Patent Information, Bayer HealthCare, Wuppertal
2005 Head of Information Retrieval, Bayer HealthCare, Wuppertal
2007 Head of Information Retrieval and Analysis, Pharmaceuticals Division, Bayer, Wuppertal
2014 Chief Scientist