The Data Advantage:
Smarter Data, Smarter Oligos

Rachapun presented the RNAHub, a multidisciplinary research community within Roche focused on advancing RNA therapeutics ― including oligonucleotides and siRNAs ― through a data-centric drug discovery approach. He outlined the in-house data pipeline used to integrate internal project data, which is typically enriched with publicly available patent data. These enriched datasets are then standardized and structured for machine learning (ML).

He also presented “OligoGym”, an open-source Python package developed by Roche to streamline the featurization, training, and evaluation of predictive models for oligonucleotide properties. Rachapun highlighted key challenges in applying ML to oligonucleotides, such as limited training data, and shared a case study on ML-guided gapmer design. 
 

Tanguy presented the in silico models used at Evotec to filter oligonucleotide sequences based on off-target effects, interspecies conservation, and binding to other targets. He emphasized the importance of incorporating diversity elements into oligonucleotide library design to support retrospective learning and improve screening outcomes. While some oligonucleotide design rules are known and understood, Tanguy noted that much remains to be discovered ― underscoring the continuous need for large-scale experimental screening to generate training data for modelling.

Building on this, Melanie introduced Evotec’s automated high-throughput screening platform, which uses RT-qPCR to evaluate large oligo libraries. Their system enables the assessment of up to 13,000 sequences per day, generating the data needed to refine and validate predictive models. 

Ming gave a preview of Genedata Screener, a component of the Genedata Biopharma Platform used by leading pharma, biotech, and CRO organizations for high-throughput screening data analysis. Designed to handle large volumes of assay data, Genedata Screener has a dedicated module for gene expression analysis — ideal for oligonucleotide and siRNA screening workflows.

She highlighted how automation features in the software accelerate and harmonize data analysis across assays and users, helping teams save time and reduce the risk of errors.
 

Matthias discussed the role of high-throughput RNA sequencing in oligonucleotide discovery and shared how Genedata is supporting this space. He also provided a preview of upcoming development plans for a new feature designed to analyze complex RNA-Seq data more efficiently.

After the presentations, the speakers joined an engaging Q&A session, where they expanded on topics raised by participants. These addressed practical challenges and future opportunities in the field, such as identifying the best tools for patent mining, using data analysis tools together with other systems such as ELNs, and using in vitro efficacy models to predict the effect of chemical modifications.