In the October 7th webinar titled “Implementing the Allotrope framework at Biotechs”, Monica Berrondo, CEO of MacroMoltek, and Dennis Della Corte, CSO of ZONTAL, offered a look at a case study between the two companies. This project used ZONTAL as an Allotrope Framework-based data management hub, enabling the use of machine learning techniques for data science. The ability to employ these methods increased the efficiency and the quality of the analysis done by MacroMoltek.
Download the key takeaways from this presentation.
MacroMoltek’s vision is to use advanced AI to shift the creation of new antibodies from a drug discovery to a drug design process. MacroMoltek is the first company to computationally design an antibody. They have a highly experienced cross-disciplinary team with backgrounds in computer science, machine learning, biology, engineering, and therapeutic development. This team uses their in-house built design algorithms, which rely on decades of structural and sequence data, to create models of neural networks that assess the target structure. They then analyze each part of the structure that was flagged by the AI as being promising for design, selecting the best lab testing models for each.
Each step of this process relies on vast amounts of data which needs to be stored, catalogued, and made available for the team to access. In this scenario, the accessibility and searchability of data is critical, especially since each team member comes from a different background. Furthermore, MacroMoltek needs a consistent pipeline, which tracks data from the earliest stages through to the validated antibodies. It is crucial to feed data back into the existing AI, so that its algorithms can continuously improve.
The challenges MacroMoltek faces in improving the efficiency and quality of their work are primarily in finding and preserving data in a way that enables cross-project analysis. To address those challenges and create a reliable data strategy that avoids the loss of data and IP, they partnered with ZONTAL, which uses Allotrope to manage information and make it easily accessible and searchable.
There are currently several schools of thought on how to leverage Allotrope in business cases, from building complex ontological graph models to ignoring all semantics and only creating an ADF as an afterthought. While complex ontological graph models might enable the most complete representation of the data, ZONTAL has found that by beginning with tabular models, they can showcase the benefits of Allotrope without overwhelming businesses with the details of the process. In this case, subject matter experts only need to give a basic idea of what might be important in a data model, after which ZONTAL can handle the data FAIRification and create ontologies on demand. Once the data is FAIRified, ZONTAL creates a self-reporting data asset and democratizes the data access by making large repositories of Allotrope files searchable through APIs, such as SQL.
Before ZONTAL, the method of data management at MacroMoltek was to run an Elisa assay, which produced one Excel document with a table representation and one plate analysis. Now, the data from multiple assays is mapped by ZONTAL and an Elisa ADF data model is created and made available for project wide analysis. This data is then preserved and accessible on the ZONTAL dashboard, where several visualizations can be created and edited to make data monitoring and analysis easy, thus addressing the challeng
es faced by MacroMoltek in preserving their data in a way that enables cross project analysis and consistent feedback for their algorithms. This case study shows that, by leveraging the capabilities of Allotrope, ZONTAL can help advance the analytical abilities of biotechs and offer all the benefits of Alltrope without overwhelming them with all the details of the process.
We invite you to watch all Allotrope Connect videos.