Anka Chan (@AnkaChan1) profile picture

Co-founder, @getruthca 🌱 Business Development @InvertBio Passionate about biotech, pharma & startups

Joined February 2022
So… is generative AI real or overhyped when it comes to drug discovery? IMO, it’s real but with a huge asterisk* As Patrick Malone is saying, the reason why AI seems like such a flop is because of our expectations around what it can actually do
one of the primary causes of confusion regarding the impact of AI in drug discovery stems from a failure to accurately communicate probabilities. fewer than 1% of drugs advance successfully through preclinical and clinical development. even if AI improves the probability of success by 2-3x (a significant improvement), the vast majority of drugs will still fail. the distinction between relative vs absolute improvement is an important one. tripling a 1% success rate to 3% is a significant relative improvement, but modest in absolute terms. the failure to distinguish between the 2 leads to inflated expectations, which we've witnessed over the last decade of AI-driven drug discovery. given the high costs and long timelines of drug development, even modest improvements in success rate can have massive economic implications and result in billions of dollars saved. but we will need to be patient. asking "where are the AI-discovered drugs?" is missing the point. considering the inherently low success rate of drug development and the incremental impact (in absolute terms) of AI on the likelihood of success, it will take time to empirically demonstrate the impact of AI on drug development.
1
1
287
We live in a world that isn’t purely bits. As much as we can simulate atoms, real-world testing still has to prove hypotheses true. AI is extremely useful in speeding up the early discovery timelines - but we have to be patient with clinical trial timelines.
1
103
Having a better understanding of clinical trials and why they stop could be a good starting place.
69
one of the primary causes of confusion regarding the impact of AI in drug discovery stems from a failure to accurately communicate probabilities. fewer than 1% of drugs advance successfully through preclinical and clinical development. even if AI improves the probability of success by 2-3x (a significant improvement), the vast majority of drugs will still fail. the distinction between relative vs absolute improvement is an important one. tripling a 1% success rate to 3% is a significant relative improvement, but modest in absolute terms. the failure to distinguish between the 2 leads to inflated expectations, which we've witnessed over the last decade of AI-driven drug discovery. given the high costs and long timelines of drug development, even modest improvements in success rate can have massive economic implications and result in billions of dollars saved. but we will need to be patient. asking "where are the AI-discovered drugs?" is missing the point. considering the inherently low success rate of drug development and the incremental impact (in absolute terms) of AI on the likelihood of success, it will take time to empirically demonstrate the impact of AI on drug development.
7
14
2
83
14,319
"If it's not on my calendar, it doesn't exist" Anyone else live by this rule or is it just me? 😂
67
21
15
370
BRB, framing this 🖼️ 👉 @AudreLawdAMercy
No we’re not raising right now
2
8
4
122
new💡from The Courage to Be Disliked, is that life isn't a competition.
1
7
I've always thought that I'd never want to live out anyone else's life, but it's becoming so real lately. Seeing friends enter new chapters of life - fresh jobs, grad school, new startups, it's inspiring to see their pursuits.
1
3
And makes me all the more grateful for my own journey, where the healthy comparison lies with my ideal self.
2
Anka Chan retweeted
You don't have to take our word for it, check out what everyone's been saying about Ruth's earth-friendly pads at getruth.ca. Did you catch us on @globalnews @GlobalEdmonton!? #periodcare #ecofriendly #earthfriendly #shoplocal #edmonton #globalnews #news
1
3
Anka Chan retweeted
5 chances to win, just by taking this survey! Did you know that 86% of menstruators get their period unexpectedly? Fill out this survey to help us make menstrual products accessible at the workplace. Survey: forms.gle/kFrkXKpniHquD2uG6 Thank you for your feedback & best of luck!
2
1
Anka Chan retweeted
Ever wonder why a period-friendly workplace is important and how to create one? Read Ruth’s article on our website to learn more! ow.ly/kxo450IyPss Make your voice heard by filling out this survey: ow.ly/62QS50IyPsr #sustainability #workplace #periodcare
2
3
twitter threads are like BuzzFeed listicles, but for "professionals" (1/2)
1
that's why I like them so much :) (2/2)
Yup @BleedtheNorth founder Isabela Rittinger said it right 👏
"Period advocacy is not radical" - Isabela from Bleed the North
2