April 22, 2019 | The Bio-IT World Conference and Expo took over the Boston Seaport area last week, as more than 3,000 researchers, bioinformaticians, and experts converged for conversations around AI, FAIR data, informatics, edge computing, and much more.
We are combing through our notebooks, gathering background, and processing all the good information we learned, but in the meantime, here’s an on-going collection on news, links, and tidbits we gathered at the meeting. We’ll be updating this as we go; please pass on anything we missed! A.P.
Bio-IT World named winners for the first Innovative Practices Awards, highlighting entries from Abbvie, Discngine, Novo Nordisk, Linguamatics, and Sentieon. An honorable mention award was given to Regeneron Pharmaceuticals and DNAnexus.
Bio-IT World also recognized the 2019 Best of Show awards, honor the best new products on the show floor. Our judges—reveling in their freedom—named a “No BS AI” award, a “Where the Puck is Going” award, a Patient-Focused Software award, and a “Nailed It” award.
The community of Bioinformatics.org also voted on the Benjamin Franklin Award for Open Access in the Life Sciences. The finalists included Richard Durbin, Aaron Quinlan, and G.P.S. Raghava. Eugene Koonin is the 2019 laureate.
Edge computing is now in the “steady development” period and AI is needed to unlock all its potential, said Weisong Shi from Wayne State University. AI is also now “easy” thanks to frameworks and systems such as Apache Hadoop, Apache Spark, TensorFlow and Alluxio. By 2020, 55% of data generated will be edge data “impossible to move into the cloud,” giving rise to “intelligent alignment software” and a shift from artificial intelligence to “extended intelligence”.
“Omics” data is helping to identify key mechanisms of Alzheimer’s disease using open-source DataLENS platform. The data is coming from sources that include the Accelerating Medicines Partnership – Alzheimer’s Disease (AMP-AD), a precompetitive partnership among government, industry, and nonprofit organizations that focuses on discovering novel, clinically relevant therapeutic targets and on developing biomarkers to help validate existing therapeutic targets, explained Sudeshna Das of Massachusetts General Hospital/Harvard Medical School. Neuroscientists are focused on publicly available gene expression data to see, for example, if a gene changes at the protein level. Genes PTK2B and TOMM40 have been found to be significant.
Can we rationalize physicians’ off-label use of drugs? Maybe. FDA’s Adverse Event Reporting System (FAERS) is unappreciated real-world data source, containing a goldmine of information (e.g., other drugs patient is taking, active ingredients and indications). Elsevier’s ResNet database is another option.
One critical component of adaptive clinical trials (aka complex innovative designs or CID) is an integrated technology platform capable of continuously collecting real-time data from the right people at the right time. At Janssen the platform called ACTIVE (Adaptive Clinical Trial’s Interactive Virtual Environment) leverages SAS Life Science Analytics Framework (systems integrator) as the primary clinical data and records repository. It manages the adaptive component of studies (such as interim analysis and study adjustments) to enable the best possible decisions at each time point, explained Vlad Dragalin, Janssen Pharmaceuticals R&D.
Kaiser Permanente is formulating its version 1.0 strategy with AI, or “machine learning” as CMIO John Mattison terms it. He pointed out systematic bias (book Weapons of Math Destruction where “opinions are expressed as code” and Automating Inequality about how AI output is only as good as the training data sets), how MD Anderson spent $62 million on overhyped IBM Watson before it was abandoned, and Alexa is listening in on our household conversations. He also believes leveraging AI is “tantamount to be a value-based care system,” but “the tool has to be fit for purpose.” In other words, the same architecture used for a round of Jeopardy can’t be applied to healthcare. The requirements for Dr. Watson are vastly different.
PatientsLikeMe has combined phenotype and molecular data to create a “research-grade” database and is aiming for “point-of-care-grade,” explained Renee Deehan Kenney, PhD, VP of biocomputing at PatientsLikeMe. It’s now grappling with the biomarker problem, hoping to eventually identify the specific treatments that might prolong the life of specific types of patients, in part by creating a prior knowledge causal network. “Genetics is one component of this massively complex and dynamic molecular system. Our focus is on assessment of the state of biology for learning.” DigitalMe Ignite launched in 2017 to follow 2,000 patients with 20 conditions but focus is on nine in three clusters (MS, Parkinson’s and ALS). The project currently has 900 patients and 1,500 samples processed and preparing to launch a giveback campaign.
Product and Company News from the Expo Floor
Genomenon announced a partnership with Google to make the company’s genomic mutation data available on Google Cloud Platform.
WekaIO demonstrated an optimized genomics data analysis workflow featuring WekaIO Matrix, the world’s fastest parallel file system; PetaGene, the maker of award-winning genomics data compression solutions; Sentieon, for accelerated genomic data analysis; and Western Digital’s ActiveScale cloud object storage for long-term storage and archival.
Rama Rao, CEO and co-founder of Bloqcube, says the company has brought together CTMS, payments, accounting all under a blockchain umbrella. We think we may be among the first in this class linking clinical trial systems and tools, he said. Rao announced a strategic alliance with NTT Data.
BioSpace’s Mark Terry has a round up, highlighting news from Altos, WASAI, Genomenon, 1CellBio and more.
Neil Versel, at GenomeWeb, covered John Wilbanks’ plenary talk on standards, open science, and the NIH’s All of Us.
Copyright Clearance Center was keeping a close eye on the tweet wall. They’ve rounded up their 19 most insightful tweets from the event.
John Pace blogged his experience as part of the Bio-IT World FAIR Data Hackathon, calling it “a great experience. I met some great people from varied backgrounds and enjoyed the camaraderie and cooperation between all of the teams. I can’t wait for next year’s hackathon!”