In this blog series, we’ve been taking a closer look at the Ingenuity Knowledge Base. So far, we have looked at the depth of content as well as the manual curation and integration efforts that make this repository unique. Today we’ll look at the advanced algorithms that compute across all of this information to deliver the most relevant and useful results to scientists using Ingenuity Pathway Analysis (IPA), Ingenuity Variant Analysis, and our new Ingenuity Clinical product.
Just as we have developed unique approaches to curating and integrating content, the team behind Ingenuity Knowledge Base has come up with sophisticated algorithms to accelerate scientific discoveries using methods you can’t find in any other web-based bioinformatics solution.
For example, while other tools can help you find links between genes or genetic elements, we have taken the next step and show you the directionality between those elements — not just that there is a link, but the nature of the linkage. Instead of “A is associated with B,” Ingenuity Knowledge Base captures the nature of the association. For example: “A increases B” or “A decreases B” (where A and B might be genes, small molecules, transcription factors, or other elements). This is very important when you want to use the Knowledge Base for computational purposes. These directional associations are leveraged by advanced statistical and modeling to generate biologically relevant predictions. This provides valuable and unique information about activation or inhibition that will set the stage for your next experiment. Such analysis is a critical component of our Causal Network Analysis and Upstream Regulator Analysis computational tools within IPA, which predict regulators, downstream effects, or whole networks based on gene expression data.
Further leveraging the computational power of these relationships, we have introduced advanced simulation tools, so applications powered by Ingenuity Knowledge Base can simulate biologically relevant upstream and downstream effects based on underlying content. Given specified activity of a particular molecule, for instance, you can see at a glance what effect that would have on other molecules, functions, or diseases. This feature, available through IPA, can be used even without your own data set, giving scientists an opportunity to explore complex interactions of molecules even before performing an experiment.
Across our algorithms, one of our guiding principles has been ease of use for people interacting with Ingenuity Knowledge Base. The greatest analytics in the world wouldn’t mean much if they were too complicated to use! We always spend a great deal of time designing the functionality, testing the usability, documenting its usage, and leveraging the computational horsepower of the cloud to run all these computations so that users don’t have to worry about processing power or elbowing neighbors off an in-demand local cluster. Taken together, these attributes make it much easier and faster for users to test a hypothesis than if they were running, say, a series of R packages, perl scripts, or some other one-off solution.
This is further demonstrated in Ingenuity Variant Analysis, where we leverage the vast array of complex directional relationships and interactions to provide biologically relevant predictions of causal genetic variation from DNA sequencing data. By using the Biological Filter within Variant Analysis, users can supply biological terms that describe the disease of study, and sophisticated algorithms will walk the ontology and compute across relationships to prioritize variants of interests. The biological terms the system accepts span from abstract to highly specific, and the semantic algorithms do the rest of the heavy lifting. This enables the user to focus on a biological area of interest while the system prioritizes variants of interest.
As we explore and expand our definition for how we think about the power of the Ingenuity Knowledge Base, Variant Analysis offers publicly available reference data sets so users can see how their results vary from normal. For example, you could select the 1,000 Genomes project as a filter within the Variant Analysis Filter cascade and see immediately that the mutation occurring at 20 percent in your population only occurs at 0.4 percent in the average population, with additional biological context provided by Knowledge Base — a real boon as you’re trying to make sense of genetic variants.
And as we push the boundaries we are discovering new utility for the Ingenuity Knowledge Base. Most recently, this includes resolving a key bottleneck faced by clinical geneticists in interpreting sequence-based diagnostic tests. In support of Ingenuity Clinical we have significantly expanded our investment and content coverage of genetic variation to disease phenotype association. As our coverage grows for each disease area, algorithms developed based on standards from ACMG, AMP, CAP, and others in the genetic testing community can help prioritize and score clinically relevant genetic variation from clinical tests. This enables medical geneticists to take advantage of the richness of sequence-based tests previously unavailable while scaling the clinical testing laboratory capacity to interpret these tests.
As you can see from this series of blog posts, we have invested a tremendous amount of time and effort building Ingenuity Knowledge Base into the underlying platform that powers the Ingenuity web applications — now you can understand why we’re so proud of it!
Thanks for taking the time to read these blog posts and learn more about the engine powering all of our great applications.