SlideIn a study with Novartis, BioXplor was provided with patient data from a phase 3 clinical trial to study patient response to anti-IL17A inhibitor, Cosentyx, versus Etanercept (Enbrel), anti-TNF inhibitor. A dataset from over 2000 patients was provided with 200 features including patient medical history, drug dosage, investigator scores, and patient genotypes. The goal of the study was to identify clinical responders and non-responders, and to identify which patient's responded better to Cosentyx versus Etanercept, based on disease stage, age, blood tests, genotype and drug dosage. BioXplor applied a custom machine learning model which successfully identified patient responders in each scenario and recommended which drug fits the right patient by combining BioXplor's proprietary knowledge discovery engine and genomics databases together with the clients clinical datasets. Additionally, SNPs linking patient adverse events to genotypes and drug efficacy were identified. Case Study - Cosentyx vs Etanercept Patient Responder ID, Personalized Medicine, Adverse Events Request More Information

SlideApplying natural language processing to annotate unstructured medical records and TensorFlow-based relationship extraction.Data was obtained de-identified from UCSF, and was in free text, unstructured format containing a 20 page medical histpry for each patient including drug dosage, side effects, bloods, demographics and duration of treatment. Insights were obtained and matched with BioXplor's broader knowledge graph on colorectal cancer to match side effects, co-morbities and drug interactions for insights into patient response. Case Study - Colorectal Cancer EMR's request more information Structuring Free Text using Natural Language Processing and Machine Learning

SlideIn a collaboration co-funded by UC Berkeley, BioXplor set out to generate novel biomarker hypotheses for development of a diagnostic assay for Pancreatic Ductal Adenocarcinoma using biomedical literature-based pathway analysis and gene set enrichment on RNAseq data from pancreatic cancer patient tissue. In this study, BioXplor identified a number of novel potential biomarkers from RNAseq data, with evidence in literature, and some others with no supporting evidence. Pathway analysis was performed to further understand these potentially novel biomarkers in pancreatic cancer. Common cancer pathway biomarkers were identified with relationships to the novel biomarkers. A protein-protein interaction database was built to identify network proximity, and RNAseq datasets from heathy versus disease patients were combined to identify 16 potential genes which could distinguish between PDAC and healthy tissue. Case Study - Pancreatic Cancer (PDAC) Request More Information Biomarker Discovery, Gene Set Enrichment, Pathway Analysis

SlideIn a study with Merck Life Sciences, BioXplor set out to apply natural language processing for the design of a multiplex PDL1 biomarker ELISA assay, comparing our designs to designs made by in house scientists manually reading papers, and those made by a in house bioinformatics team. BioXplor employed a literature discovery approach to establish a pathway model to prioritize PDL1 related proteins to measure in an ELISA assay for PDL1 activity in drug screening. Then, BioXplor's team extracted the optimal antibody reagents, prioritizing based on similar concentration, tissue-specificity, secondary antibody combinations, and target specificity supported by the highest ranking literature citations. BioXplor successfully identified key biomarkers missed by the human scientist and bioinformatics team, and produced superior results for antibody selection further improved upon the existing commercial assay. Importantly, BioXplor's designs were performed 20x faster than the human scientist outperformed the bioinformatics team in biomarker identification. This approach is highly scaleable, and similar results can be obtained in your workflow via integrated of BioXplor's API or our software tools. Case Study - PDL1 ELISA Assay Design request more information Biomarker Discovery, Pathway Analysis, Antibody Reagent Mining