Big Data and personalized medicine

Question 1

Medical data

- from data to wisdom

Answer 1

data: collected raw material, no connections, no relationship
information: organized & structured data
knowledge: give information certain meaning
wisdom: integrated knowledge
decision: purpose

Question 2

form data to wisdom

- conversion zones

Answer 2

found between the levels

• conversion form one level to next
• can be manually or automatically done

Question 3

big data drivers

Answer 3

• technological advances → imaging & sequencing techniques
• IT innovations → IoT, wearables
• Digitalisation → elecronic health record
• social behaviour moments → measure and share information

Question 4

big data problems > 4 Vs

Answer 4

• Volume: data in scale
• Variety: data in many forms → strucured, unstrucutred, format
• Velocity: data in motion → analysis of streaming data
• Veracity: data uncertainty → managing reliability and predictability

Question 5

Medical data

Answer 5

medical images, genetic profiles, questionnaire, metabolic & mobile monitoring
- analytics: needed to connect all data together

Question 6

challenges of big data analysis in health care

Answer 6

• data can be inconsistent
• analytical issues
• problems of integration into clinics

Question 7

database definition

Answer 7

collection of organised data

• Softwares: Database Management System DBMS & MySQL
• created to store data → medical laboratory results and images

Question 8

relational model of data

Answer 8

• data stored according to certain models in a DBMS
• Relational model most commonly used
• data organized in tabular format with rows and columns

Question 9

SQL

Answer 9

• SQL = structured query language
• computer language for interacting with relational database
• special purpose language

Question 10

SQL commands

Answer 10

Data definition language
- CREATE TABLE, ALTER TABLE
Data manipulation language
- INSERT, UPDATE, STORE
Data control language
- GRANT, REVOKE
Transaction control language
- COMMIT, ROLLBACK

Question 11

Big data and relational DBMS

Answer 11

Problem: scaling

• scale up → vertical scaling (larger computer for more data storing)
• scale out → horizontal scaling (more computers for more data storing)

Question 12

Data lakes

Answer 12

Idea: add all data in a “lake”
not strucutred, raw data
challenge: need different models

Question 13

noSQL

Answer 13

= not only SQL
- Key-value: volatile or persistent
- Wide column → ID
- document → describe data by different means
- graph → connect data points together
→ advantage: more room to organize data

Question 14

role of big data in personalized medicine

Answer 14

extract form world population medical data and day-to-day data → store them → via data analysis new knowledge can be formed → specifically target patients with certain treatments

Question 15

personalized medicine

Answer 15

medical problem can be viewed as emergent structure and cells, tissues, molecules are the individual interactions

Question 16

medical data variety

Answer 16

problem: medical data, imaging etc → difficult to connect
- solution: computing framework
→ Variety (combining existing models) & Veracity (use simple discrete models)

Question 17

Example for personalized bone health prognosis

Answer 17

cellular automation
- discrete domain → tissue
- cell status → tissue density
- local update rules → mechanical and molecular regulation
boolean network
- boolean variable → molecular/cellular states
- boolean functions→ molecular/cellular interactions
Prototype
- patient specific input data
Simulation → bone loss related to adolescent idiopathic scoliosis

Question 18

SQL vs mySQL

Answer 18

SQL: used to manage relational databases
mySQL: used to store, retriev, modify and administrate data using SQL