As the newest member of the Applied Machine Learning team based in Boston Sara is excited to meet with you and brainstorm to define projects and form collaborations. She is a Bioinformatics and Computational Medicine researcher, expert in integration, analysis, visualization, and interpretation of (pre)clinical and high-throughput data. Her work involves developing algorithms to apply machine learning, heterogeneous data integration, and, 'omic' data and sizeable networks analysis to study complex genetic diseases (e.g., cancer and autoimmune disease), therapeutics response prediction, and computational functional genomics. Sara has over fifteen years of experience working and presenting in highly multidisciplinary teams of MDs, computational scientists, and biologists., and, writing articles and grant applications.

I am a pharmacist from Rio de Janeiro-Brazil and I work as a computational chemist at Bayer since 2021 in the Computational Molecular Design department in Wuppertal. I work side by side with chemists on different projects using our digital tools to design small molecules that maybe one day will help patients! I am passionate about my work and I have the opportunity to also be part of the Steering Group for Transformation and Leadership in the Drug Discovery Science to help transforming the way we work and communicate.

Understanding how compounds affect the disease biology of patients, and which potential side effects could be expected, is one of the core challenges in drug discovery. Transcriptomics (RNA-seq) allows us to measure these compound effects via an unbiased readout of gene expression changes. While „conventional“ analysis of RNA-seq data yields insights into individual compound effects and similarities, deeper connections between principles of compound structure and transcriptome effect elude us. Insights into this connection would help us support project teams with data-driven decision making on compounds. In this pilot, we are bringing together a cross-functional team of colleagues from LST, CMD, SyMOL and DS&AI to tackle this challenge and build a toolkit to enable understanding of ​how compound structure relates to complex gene expression patterns​ and to explore common themes amongst structures or shared off-target/general perturbation effects. This toolbox will become the starting point for other efforts aimed at leveraging multi-omics data in understanding compound mechanism of actions. Artwork created using Stable Diffusion.

The SCLM4Future project together with the PRINCE (Preclinical Information Center) project as well as the CONDAS (Connecting Data for Science) team is an excellent example for a successful cross-functional collaboration between CD&O, RED preMed and DS&AI within Research & Development but also between Platform IT and external partners. Together the team achieved to overcome legacy manual processes, redundant maintenance work, and functional silos by introducing new digital solutions and concepts with a strong collaborative mindset. The newly released SCLM 2.0 (Standard Codelist Maintenance) & PTO (Preclinical Terminology and Ontology) systems will enable our organization to improve the quality of controlled terminologies with regards to standard codelists and codes while reducing maintenance time and effort. This will have a positive impact on the code/codelist consistency across clinical and preclinical, as well as the compliance to regulatory requirements (e.g. CDISC (Clinical Data Interchange Standard Consortium) Controlled Terminology). The approach to share our knowledge as well as our efforts across functions will be an important step to archive our goals to increase data quality, interoperability and reusability – or to make a long story short: to make our data even more FAIR. For more information contact: Daniela Bergann.

Hubble is a pioneering data hub that continuously integrates and enriches terabytes of external scientific textual data to support insight generation along the R&D value chain. It opens opportunities for statistical analyses, large-scale text-mining and other data science methodologies to tease out important details and detect trends within e.g., patents, scientific literature, and grants. The Hubble platform is available Bayer-wide, and offers API access for data scientists. This allows large scale, programmatic analysis as well as integration into processes and platforms. Hubble is a cornerstone of our digital transformation journey within R&D. It serves many users, systems and processes with key data & analyses by answering ~2 million queries per week. We continue to extend the platform to best support insight generation and scientific decision making with a precision medicine focus. For more information contact: Astrid Rheinländer.

Scientific work in Pharma R&D requires a reliable high-quality body of data and knowledge. Bayer Pharma R&D has an extremely large amount of scientific data. The use of a Data Catalog significantly improves the findability of data and serves as a key pillar in the implementation of a comprehensive data strategy.

Over the past two decades, the interdisciplinary predicTeam has established a prediction platform at Bayer Pharma R&D with the goal to generate state-of-the-art machine learning models for a variety of pharmacokinetic and physicochemical endpoints in early drug discovery. These tools are accessible to all scientists within the company and can be useful in assisting with the selection and design of novel leads, as well as the process of lead optimization. The predicTeam provides an all-inclusive package covering the data pipeline from experiment to application in projects. In close interaction with experimentalists, we select endpoints for model building that are relevant for drug discovery. We implement and maintain the infrastructure to retrieve and prepare the data and make it accessible as a data lake. For each endpoint, after fully exploring the matrix of data, molecule representations and algorithms, we implement the best-performing and most stable-models models in our internal research platform Pix. A highly automated infrastructure allows us to perform weekly retraining of the models to ensure that the novel chemical space of drug discovery projects is well embedded. We ensure close interaction (e.g. presentations, tutorials, teams channel) with the user base for optimal model use and direct feedback allowing for constant improvements. Finally, our Model Performance Report helps users to assess the applicability of each model to their specific project molecules.