With the advent of high throughput screening and next generation sequencing, pharmaceutical and biotech companies have massive amounts of information to identify clinical patient response to drug therapies. This data is high-dimensional and complex with subtle substructures, and may include gene expression data, clinical data, and third party data sources.
While biologists, geneticists, and clinical data practitioners have a wealth of data available to them, conventional methods, such as principal component analysis, multidimensional scaling, and hierarchical clustering, fall short. The Ayasdi Platform provides a new approach to complement existing techniques to significantly shorten the drug discovery process, resulting in faster time to market -- saving millions to billions of dollars along the way.
Ayasdi's use of Topological Data Analysis (TDA) is ground-breaking, and Ayasdi's technology makes it easy for anyone to mine and discover insights in large, multi-dimensional data sets to drive rapid innovation in clinical trials and drug discovery.Amanda Enstrom
Getting an accurate picture of genomic diversity and similarity in human populations constitutes a huge step in understanding diseases. Similarly, understanding the genetic causes for a specific disease and unlocking its gene expression signature in a homogenous population gives a profound insight into personalized treatment approaches.
Next Gen Sequencing
Using the WTCCC (Wellcome Trust Case Control Consortium) dataset, IRIS identified a cancer-related genomic variant that stratifies the Japanese and Chinese populations, their associated gene expression profiles, and specific subtypes of gastric tumors.
Download the Poster: Next Gen Sequencing with TDA Analysis
Copy Number Variation Analysis
Using copy number variation (CNV), IRIS distinguished three populations (East Asian, European Americans, and Yorubas) in the HapMap data set (subset of the 1000 genomes project). Within minutes, familial relationships were identified between samples in this extremely large dataset.
Download the Paper: Analyzing Population Structure Using CNV
Mapping DNA Microarrays
Using Microarrays, IRIS identified a unique subgroup of Estrogen receptor-positive breast cancers that exhibit 100% survival with no metastasis. These insights offer new methods of unlocking potential of streamlined and personalized medicine.
Download the Publication: TDA Identifies New Sub-Group of Breast Cancers
Analyzing complicated data sets from heterogeneous sources, such as mapping gene combinations across experiments, is a big data combinatorial challenge. The ability to use data that facilitates an understanding the biological and chemical mechanisms behind a disease can lead to better drug targeting and treatment approaches.
There are many biological and clinical cases for which patterns and causes are not clear, resulting in incomplete patient diagnosis, therapy and treatment. The ability to use data to move beyond classifying patients into overly-simple categories is crucially important to to develop innovative and effective strategies for patient care and treatment.