Transforming Big Data Into Meaningful Knowledge

Systems Imagination Inc. (SII), specializes in transforming big data into meaningful knowledge. Our computer scientists, mathematicians, and biologists can answer questions like – how can we help our customers identify potential diagnostic and prognostic markers to detect diseases like cancer more quickly and earlier than ever before? To do this, one must examine the complex biological networks and entities involved in a disease’s progression – everything from proteins, to metabolites, to entire genomes, to tissues, to lipids and a seemingly endless myriad of biology to understand the disease. All of these biological elements must be represented and linked together despite different data types, levels of complexity, and functional understanding. Our job is to integrate the data in a consistent and comprehensible way. One way to represent this harmonized data efficiently is with a hypergraph database.

Hypergraphs can be used to model broad connections among elements of big datasets that make more sense when categorized together. The healthcare and life science data that SII works with has multiple levels of information, and a variety of perspectives are needed to grasp these complexities. Biological systems often only make sense when we can comprehend many elements working together. Hypergraphs could help accelerate our understanding of the processes occurring by better representing the multiple functional relationships between biological components. Protein complexes, for instance, behave in groups, yet most of the protein-protein interaction (PPI) databases today represent interactions as an edge connected by two vertices. This representation makes it challenging to detect for similar functional modules, which is essential to understanding the behavior of proteins in the body. Further, multiple molecular interactions in a given biological pathway often synergize such that their interactions only seem to make sense in combination with each other (i.e., the reactants and products are a directed relationship such as A+B-> C+D, where A, B are the reactants and C, D are the products). Specifically, a directed hypergraph would be a closer approximation of the processes occurring in complex molecular mechanisms in a cell.

Hypergraphs have been explored in the context of knowledge representation and graph- object-oriented databases, and database schemas have (sometimes inadvertently) used a hypergraph-based framework. However, there isn't a commonly used hypergraph biological database, likely because biological networks are complex. Fusing that data together in a way that preserves biological meaning from standalone datasets can seem even more challenging. Systems Imagination is up for the challenge, and we enable our customers to find answers faster using our hypergraph databases.

For more information on graph databases check out O'Reilly's Graph Databases book .

About Kendyl: Kendyl is the Applied Mathematician at Systems Imagination.  Experienced in the development and use of deterministic and stochastic models of biological processes, Kendyl has worked with customers to structure massive biomedical data sets into computable knowledge graphs to help scientists identify meaningful insights.

Read More in my Whitepaper: Hypergraphs and Database Implementation

Posted on

September 4, 2018






Years of Experience


Positive Feedback


Sleepless Hours


Cups of Coffee