Marc Tessier-Lavigne
speaker
407 appearances
1 recordings
1 series
first heard Mar 2026
last heard 12 Mar
Marc Tessier-Lavigne’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Mar 2026 with 1.
Appearances
What causal actions are necessary for that?
Those models fail completely.
And the reason they fail is because they're trained on descriptive data.
So you train on descriptive data, you're good at descriptive questions.
You fail at causal questions.
What's needed are causal data.
There are various ways of creating causal data.
One of the ways we're doing it is using a technology that enables you to modify in cells one gene at a time.
So switching off genes.
So you take a cell, you might take, as we've done, a liver cell or immune cell or a cancer cell.
and switch off each of the 20,000 genes individually in those cells and read out what's the effect on all 20,000 genes.
So it's a 20,000 gene perturbation experiment by 20,000 gene readout.
Those are the kind of data, causal data, that are necessary to create a model that starts to understand the genetic networks in the cell at a causal level.
So we have published and we are generating massive amounts of causal data to train causal models of the cell to be able to ask that question, what does it take to go from a disease state to a healthy state?
And we believe that those kinds of data
will be sufficient in the right quantities and the right cell types will be sufficient to be able to address questions that are useful therapeutically.
Again, just to answer your question, do we need to have a complete model of the cell to be able to make progress?
The answer is no.
We believe that this will be a first draft version of the cell, if you will, first draft causal model of
on which we're going to want to layer much more data.
Showing 201–220 of 407 · page 11 of 21
← Previous
Next →