Client success story
ML Driven Target and Biomarker Data Mining
Business Value: Robust Oncology Biomarkers Identified
Summary: Machine Learning On Multiple Assembled Data Sets
An early-stage oncology biotech company is developing early cancer detection tests in conjunction with potential therapeutics, but the data needed to identify target pathways is complex and diverse. bPrescient assembled data sets and built analytical tools to allow the client’s scientists to identify useful subsets of data that collect a variety of data types from internal and external sources.
Challenge: Data Associated With Tumor Types Complex And Disparate
The client had found a set of genes of interest for their diagnostic panel but wanted to assemble more information about correlation between the panel members and individual cancer types, and identify underlying biological signals that could point to pathways or drug targets of interest. The data required to perform this analysis was varied and heterogeneous, so the client’s scientists were finding it difficult to ask the right question and get answers efficiently.
Solution: Identify Correlations In Integrated Multi-Modal Data Sets
bPrescient’s team reviewed available and assembled relevant data sets from public sources and linked them to the client’s internal data. The team then applied machine learning techniques to identify correlations between the variant list and co-located genes and proteins. Those genes and proteins were analyzed to identify pathways of interest in oncology therapeutics and led to hypotheses about targets that could be pursued and correlation between variants and disease severity in different indications.
Outcome: Biomarkers And Targets Identified
The client was able to organize the relevant data to ask their biological questions and identify potential biomarkers and targets that formed the basis for their next round of diagnostic and therapeutic development.
Let’s talk
Tell us about your toughest R&D challenge and let’s discuss how bPrescient can solve it.
