“Translation and Convergence for Future Medicine”

• 미래 의학을 위한 중개 및 융합연구
• Asan International Medical Symposium 2016
• Innovative Future for Medical Science & Technology -
• 2016년 6월 17일 (금) 서울아산병원 동관 6층 대강당 외 AIMS

Plenary Session II “의료기술및 R&D 변화의최신동향”

• Chairperson :김청수(서울아산병원비뇨기과교수)

13:30 ~ 14:15 Lecture 1 :의료분야에서의 빅데이터 :임상연구 및 진료를 위한 애널리틱스의 활용

• Speaker: David W. Bates (Harvard University, Brigham and Women’s Hospital, USA)
• Rising costs
• Moneyball, Boston red sox, walmart, watson
• Big data 1M - 1giga(human genome) - 1 peta
  • EHR, Genetics, Diagnostics, Mobile devices,
• Meaningful Use - EHR - growing.
  • https://www.healthit.gov/providers-professionals/meaningful-use-definition-objectives
  • https://healthit.gov
• Big data concepts
• Validation is important!
• Big data and research - Brigham and Women’s - Pathology ePath, Immunology Big data Genomic platform
• Essential for future approach
• RPDR - New entity at partners healthcare = CMS, biobank, survey data, imaging, notes repo
• Big data in clinical care
• 5% patients ~ 50% cost
• iCMP claims-based approach - 3000 patients
• multiple parameters - wearable devices - continuous supervision on general care floors
• Adverse events
• PCORnet - not popular
• New Sources - the trajectory of mobile apps
• Literature Review - 7301 titles and abstracts
• App Review - iTunes, Google Play -> possibly useful 16
• Professional Society Review -
• Ginger.io https://ginger.io/
  • to drive better health outcomes through the use of passive mobile data and behavioral analytics.
• !!! Example projects - Predictive Modeling
• What we need to do all these
  • Analytics tools, repo, data warehouse, epic reporting (Clarity reporting database)
• Clinical data - ubiquitous
• !!!!Novel sources are most likely to provide marginal improvement - social, mobile!!!!
• Predictions / Implications
  • Transformative as the internet
  • Killer app - Google Maps

Questions

• 김규표 교수님 - Social media and health care
• 김청수 교수님 - Government and insurance - reasonably difficult to acquire - costly.

14:15 ~ 15:00 Lecture 2 :합성 항체에서 합성 단백질로

• Sachdev Sidhu (University of Toronto, Canada)
• The Donnelly Centre - From systems biology to systematic treatment
• Therapeutic antibody revolution - highly versatile, numorous diseases
• Ab-durg conjugates, fragment, bispecific, engineered cells
• Targeting cancer with antibodies
• !!! problem - small populations - boutique treatment!!!
• In vitro protein evolution
• Affinity, specificity enhanced
• Antibody molecules
• binding site of Ab
• highly optimized - automatic mutation of binding site
• Toronto Synthetic Antibody Library - highly diverse - Herceptin
• PHAGE - Genentech
• only changed the function, not others
• Functional genomics - Large-scale, industry-quality Ab generation - Preclinical biology | The middle was not quite available but now it’s doable.
• Cancer Antibody TRAC antibodies - bacerial pathogens
• High yield and high affinity Fabs from naive library - 1394 total against 80 targets
• http://sites.utoronto.ca/sidhulab/about.html
• natural - synthetic Ab - synthetic proteins - synthesizable proteins
• D-protein therapeutics.- samll proteins synthesized entirely rom D-amino acids, Ab like affinity, specificity and stability,
• in vitro d-protein advantages - longer circulating HL than L-proteins - less immunogenicity - ersistant to metabolism in plasma

Parallel Session I “의료분야에서의빅데이터” Chairperson :김태원(서울아산병원임상의학연구소장) [대강당]

15:20 ~ 18:00

Lecture 1 :전자의무기록에 기반한 임상 빅데이터 연구 Alexander Turchin (Harvard University, Brigham and Women’s Hospital, USA)

• Data warehouse : integrates data from multiple sources - i2b2 | ABLE | OHDSI
• who entered the data? Wrong input to public repo (DKA for 2 years!)
• Data quality
• Raynaud’s syndrome - Omega3 (Failure)

Lecture 2 : 의료분야에서의 빅데이터 분석 Tom Lawry (Microsoft Corp., USA)

• 8 seconds = Concentration time
• Analytics Convergence Zone - Clinical data, Geo/Social/Environmental data/Claims&Cost Data/Pharma&Life Sience Data, Patient & Citizen Data
• https://powerbi.microsoft.com/ko-kr/ !!!!! 반드시 사용해 볼것. 좋은 Visualization.
• http://www.ciokorea.com/news/29118

Lecture 3 :생물기작 기반 암 오믹스 데이터 분석 기법

• speaker: 김 선 (서울대학교 생물정보연구소)
• https://sites.google.com/site/biohealthinformaticslab/sun-kim
• http://bioinfo.snu.ac.kr/main/index.php
• DNA, RNA, Protein이 중요 - Somatic mutation뿐만 아니라
• Transcriptome (RNA-sequencing data)
  • 싸고 쉽다.
  • 그에 비해 Underestimated되어 있다.
• Breast cancer
  • 가장 잘 알려진 암종.
  • 21-gene Oncotype DX !!! http://www.oncotypedx.com/
  • PAM50 - Prediction analysis of microarray by 50 gene classifier !!! - Survival 예측하는 Gold-standard
• Transcriptome Data analysis http://prosigna.com/x-us/overview/
• Pathway (context) analysis는 과연 informative한가?
  • A Critical Evaluation of Network and Pathway-Based Classifiers for Outcome Prediction in Breast Cancer (PLoS One 2012) http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0034796
  • 50개를 랜덤하게 취해도 유의미하게 나왔다. 아무거나 취해도 유의미하게 만들 수 있다. (Negative result!)
  • “Based on these results there is currently no reason to prefer prognostic classifiers based on composite features over single genes classifiers for predicting outcome in breast cancer.”
  • 따라서 50개를 정할 때 기능적인 면을 고려해야 한다.
• PPI-based Pathway Decomposition !!! - 우리의 분류를 기반으로 Survival plot를 그리면 확연한 차이가 난다.
• Decomposed pathway and its activity measurement Using RNA-sequencing data
• 이러한 알고리듬으로 RNA-sequencing data 분석 (1138개의 sample을 사용)
  • TCGA data (Breast Cancer)
  • 기본적으로 information theory를 응용한 것이다.
  • Subtype과 발현량의 ranking에 따라 score를 매긴 후 distinguishing할 수 있다.
  • 클래스마다 차이가 확연한 Subnetwork A는 좋은 모델
  • 클래스마다 차이가 없으면 Subnetwork C는 나쁜 모델. 우리는 좋은 모델을 택해야 한다.
• Top10 Regulated pathways by TF/miRNA
• mir-30a (basal cell cycle activation) -> mir-149, let-7b, mir-30a
• Sub-network mining approach
• Experimental validation requires collaborators who enjoy new approaches.
  • 혹시 talk를 들으시고 관심있으면 Contact.

http://www.amc.seoul.kr/asan/academy/event/eventDetail.do?eventId=572 http://aims.amc.seoul.kr/asan/depts/aims/E/deptMain.do