Overview
DNAlytics serves the healthcare industry with data sciences expertise, supporting biomanufacturing companies in optimizing yield and quality, hospitals in addressing regulations and improving data interoperability, and R&D teams in drug development and clinical research.
Key people
In the news
- LiblineaR, the R package our founder Thibault Helleputte has contributed to maintain since 2010 (even before company incorporation), just shipped its most substantial maintenance release in years (2.10-25). It wraps Prof. Chih-Jen Lin's LIBLINEAR C/C++ library for fast linear classification and regression. 🏅 Over 574,000 downloads since its first release, placing it among the top ~6% of CRAN's 20,000+ packages by downloads (per datasciencemeta.com). Most of its users never think about it, which is exactly the point of good
- Insights, a model, and a recommendation, from a biomanufacturing dataset that's already there, in one sitting, no data sciences team required. 🔎 That's what our latest Hercule video walks through: the PIE workflow (Process Improvement and Exploration) end to end, screen by screen. It's built on one principle we care about: Hercule accelerates your process experts, it doesn't replace them. Every step surfaces evidence to weigh, not a verdict to follow: a correlation worth investigating, a model worth pressure-testing, a
- Bootcamp UCAD/AFRETEC 2026 — Bioinformatics, AI & ML in Life Sciences : mission accomplie. 🌍 En juin dernier, plus de 40 participants de toute l'Afrique ont suivi l'édition inaugurale de ce programme intensif, en ligne, depuis l’Afrique du sud, l’Angola, l’Égypte, le Kenya, le Maroc, le Nigeria, le Rwanda et le Sénégal, suivi d'une semaine en présentiel au centre de recherche en génomique appliquée du CIGASS à l’ Université Cheikh Anta Diop de Dakar (UCAD). Neuf demi-journées de formation théorique, huit demi-journées d'exercices
- Ask a biomanufacturing team how often they run a real process-improvement study. The honest answer is usually: "rarely". Once every couple of years, if it happens at all. 🤔 It's tempting to read that as a lack of ambition. It isn't. And it isn't a lack of skill either. It's the data. 🤯 Before anyone can analyse anything, someone has to pull the data out of LIMS, ERP, historians and paper forms, clean it, align it, and make it audit-ready. That alone can take weeks or months. By the time it's ready, the moment has often
- Continuous Process Verification isn't a feature you buy. It's a commitment you make to your regulators, one you have to honour, again and again. 🫵 You define the plan: which parameters to watch, which rules apply, how often to review. That part should always be yours. ⚠️ The problem is execution. Done by hand, gathering the data is slow and hard to keep audit-ready; computing the statistics with scripts is hard to reproduce, review after review. Weeks of work a year, for a result that's often still suboptimal. Our new video shows
- Before CPV, before any model, before a single batch record, one thing has to be right first: how your process is actually described. It's the least glamorous step in biomanufacturing data. It's also the one everything else depends on. 🎬 Our latest video shows how teams now do this themselves in Hercule. With the new Process Builder (released in v1.4), you define your process directly on the platform: the full ISA-88 hierarchy of unit procedures, operations, phases and measures, in minutes, without spreadsheets and without waiting
- Continuous Process Verification is mandatory — and for many teams it is one of the most resource-intensive activities they run: often manual, and limited in the number of parameters it can realistically cover. At A3P Belgique a few weeks ago, our colleague Marie Bastin made the case for turning CPV from a regulatory hurdle into a source of value. The shift is twofold. 1️⃣ Automation: what takes weeks of manual work, such as control charts, capability indices, trend and normality checks across dozens of parameters, can run
- Biomanufacturing runs on data — and yet, for most teams, that data is scattered, siloed, and impossible to actually use. We made a short video about what it looks like when that changes. It's the first in a series introducing Hercule, our platform for turning biomanufacturing data into something you can finally rely on: "Biomanufacturing data, finally under control." We filmed it at the EU Biotech Campus in the BioPark Charleroi : a fitting backdrop, at the heart of the biotech ecosystem we work in every day. Over the coming
Related profiles
Something wrong or missing? Send an update. Fixed within 24 hours.