This week, CCC was among 3,400+ life science, pharmaceutical, clinical, healthcare and IT professionals from over 40 countries at the Bio-IT World Conference & Expo at the Seaport World Trade Center in Boston, MA.
If you couldn’t attend Bio-IT World’s annual conference in Boston this year, have no fear. We’ve rounded up some of our favorite bite-sized tips, session snippets and inspiration from the show:
At #BioIT19 Vijay Bulusu asks the audience if they had $1M would they improve the quality of the data or buy a machine learning platform. Nearly everyone would improve the quality of data.
— Cambridge Innovation Institute (@CIInstitute) April 18, 2019
AI panel at #bioit19 notes that the last decade focused on getting more data. Next decade will be about what to do w all that data . #collaboration @PistoiaAlliance #AI
— Carmen Nitsche (@cnitsche) April 17, 2019
“Genomics is at risk of becoming the sport of kings, where only the wealthy can play.”
— Anthony Philippakis on NGS data growth/analysis challenges @broadinstitute #BioIT19— Kevin Davies (@KevinADavies) April 18, 2019
I want to spend my time doing data science not cleaning data – that’s like asking a surgeon to get their instruments and operating room ready #BioIT19
— Anne Deslattes Mays, PhD 🇺🇦 🇭🇷 (@adeslat) April 17, 2019
In the life sciences, we don’t have a big data problem. We have the “lots of small data” problem, says Bulusu. #BioIT19
— Bio-IT World (@bioitworld) April 18, 2019
Dana Vanderwall: To successfully achieve digital transformation of the lab, it needs to be automated, self-documenting, and interoperable #BioIT19
— Kaitlin Searfoss Kelleher (@KaitAKelleher) April 17, 2019
#BioIT19 what does a good data strategy look like? @AstraZeneca pic.twitter.com/cFt4ChDjjn
— Imran Chaudhri (@imrantech) April 18, 2019
Vijay Bulusu: @pfizer Converting data to insights — while there has been an explosion in machine learning #ML and big data #BigData, there has been a fundamental lack of progress and investment in improving the “quality” of data #focusontherightthing #BioIT19
— Ken Stineman (@KenStineman) April 18, 2019
Amazing keynote at #BioIT19 by @wilbanks – open science is enabling high trust distributed network of communities. pic.twitter.com/hMvdYqFBPc
— Rachael Acker (@rachaelacker) April 17, 2019
Open Data Report – #BioIT19. 73% agree that data sharing is a good – are scientists equipped to FAIRylify their data
— Anne Deslattes Mays, PhD 🇺🇦 🇭🇷 (@adeslat) April 17, 2019
“Math, data and computation are powerful ways to figure out what we don’t know.
Think about the black hole [image]—it’s time we did that in biology and healthcare.”
—@IyaKhalil #BioIT19— Kevin Davies (@KevinADavies) April 17, 2019
Anne Carpenter – need more user friendly interfaces to AI to allow average biologists to use the tools #BioIT19
— John M. Greene, Ph.D. (@jmgreene83) April 17, 2019
Data (clean, usable, accessible, with metadata) is still the biggest barrier to progress in #AI. #healthdata #BioIT19 pic.twitter.com/ThukY4IGgQ
— Janice McCallum (@janicemccallum) April 17, 2019
AI in practice: How do you know you are not perpetuating bias in the machine learning model? Bias is inherent to novel discovery; include subject matter experts, iterations, and discipline around the metrics, constraints, and limits of the models #BioIT19
— Ken Stineman (@KenStineman) April 17, 2019
*Usability* is an important advt of AI. "AI will identify cells [in microscopy image] without biologists tweaking knobs. We’re almost there… Biologists carrying out image analysis without knowing anything about image analysis is really exciting."
— @DrAnneCarpenter #BioIT19— Kevin Davies (@KevinADavies) April 17, 2019
Keynote panel on "AI in Practice" #BioIT19 "Usability is key" "#AI #ML is a requirement for analyzing all that data – eg to better understand why still 80% of CTs fail – maybe we might get it to 50%" – Interesting polls knowing there are lots of experts here … pic.twitter.com/aYGTEntInN
— Hans Constandt (@hconstandt) April 17, 2019
Doesn’t data need standards? Have to get the Core data bits to learn from data. W3C standards RDFs, DCATs, use of persistent and unique identifiers. Contextual metadata is needed. Instruments needs to produce data with appropriate metadata #linkeddata .@jacksonlab #bioit19
— Anne Deslattes Mays, PhD 🇺🇦 🇭🇷 (@adeslat) April 17, 2019
Christianson: theme to digital transformation of R&D: adoption of digital tech; real-time data availability; advanced analytics & predictive science #bioit19
— Kaitlin Searfoss Kelleher (@KaitAKelleher) April 17, 2019
The need for open science @wilbanks:
“We’re taking tiny fragments of science and locking them up as PDFs.” #BioIT19— Kevin Davies (@KevinADavies) April 16, 2019
Related Resources: