L5: AI in biotechnology Flashcards
technologies alone can transform healthcare. T/F?
False
challenges of innovation [4]
- high cost
- long development time
- low success rate
- difficult to determine which drug and dose to use
between development and discovery, which stage takes up a longer time? state what is done during this stage.
discovery.
identify drug target, generate leads, optimise them
how can AI help in the discovery of new drugs?
- aid in drug discovery: target validation and lead optimisation
- drug development during pre-clinical and clinical: help find out the best combinations of drugs and doses [optimisation]
- N of 1: personalised therapy: optimise the duration of treatment and drugs according to patient
- digital medicine: combat cognitive decline by having programs for patients
what is N-of-1 Healthcare?
to use a patient’s own data to manage their own treatment
identifying synergy alone is enough for optimisation. t/f?
false. intra-patient drug synergism and antagonism is dose dependent
why do clinical trials fail so often?
drug synergy and antagonism needs to be re-optimised AT EACH STAGE OF DEVELOPMENT to de-risk drug development
how can we use AI to globally optimise drug development?
- finding the optimum combination of drugs from a massive drug-dose space
- drug interaction analysis: can find unforeseen interactions for evolving drug sets
differentiate between traditional and current small data AI-based optimisation
Traditional:
- based on pre-existing data
- FIXED dose
- find the best option out of n number of combinations
Small data:
- less data points [still have diff combinations]
- but have variation in dose –> find best combination out of n no. of assays representing 1 trillion++ possibilities
- use actual expt results, not pre-existing data
Perks/use of using Identif.AI
- find best permutation of drugs and dose
- ranks the combination–> know which combi to avoid
- dynamic analysis wrt to patient’s own progress and reaction to drug
how can AI make healthcare more efficient and affordable? [4]
- guide treatment choices -> personalise medicine for patients
- more efficient diagnosis [lesser time and cost]
- improve clinical trials optimisation and drug development
- empowering patient