A best, short, and clear illustration of statistical calculation like mean values and root mean square values, molecular dynamics, and free energy. I hope you have an equivalent lecture on molecular vibrations, vibrational frequency, relaxation time, and equations for drug Kinesis. Thank you.
can this also be an explanation for why docking simulations are able to produce some enrichment without being able to calculate the absolute value of kd??
Can you make a video about absolute FEP? As you mentioned, the computing power keeps increasing, so at some point it would be better to go to ABFE calculations for hit drug discovery campaigns, don't you think?
Indeed, and if it came across as hinting we could sample it exhaustively that wasn't my intention. However, when compared to the low-throughput methods such as anything experimental or detailed free energy calculation, the throughput of docking is, to first approximation, infinitely higher. In practice I would argue the problem is a different one, though: First docking is not particularly accurate, second it's not that hard to find hits, and third it would be WAY to expensive to synthesize thousands of random chemicals from scratch just because they happened to be a good hit in docking. In practice, docking is thus typically used either with libraries of compounds that can be bought, or compounds that can be synthesized cheaply by combining a few functional units. In this setting the computational cost of docking is not any limiting factor - and once we have a hit the work is rather focused on refining it (because we don't need the best possible hit in universe - just something that works).
@@eriklindahl I don't disagree with what you say. It's just that in the video, your justification for moving to "higher accuracy" FEP calculations is because "you ran out" of molecules to compute using docking. I have a problem with this rationale. I feel the video would have been more accurate if instead you said, that we now move to a higher theory, first principle approach, which is limited in scope but more accurate than docking.
A best, short, and clear illustration of statistical calculation like mean values and root mean square values, molecular dynamics, and free energy. I hope you have an equivalent lecture on molecular vibrations, vibrational frequency, relaxation time, and equations for drug Kinesis. Thank you.
can this also be an explanation for why docking simulations are able to produce some enrichment without being able to calculate the absolute value of kd??
Can you make a video about absolute FEP? As you mentioned, the computing power keeps increasing, so at some point it would be better to go to ABFE calculations for hit drug discovery campaigns, don't you think?
We're already going there. I might be able to say more in 6-12 months.
The chemical space is so vast that we can't test them all using docking. The introduction is a bit misinformative.
Indeed, and if it came across as hinting we could sample it exhaustively that wasn't my intention. However, when compared to the low-throughput methods such as anything experimental or detailed free energy calculation, the throughput of docking is, to first approximation, infinitely higher. In practice I would argue the problem is a different one, though: First docking is not particularly accurate, second it's not that hard to find hits, and third it would be WAY to expensive to synthesize thousands of random chemicals from scratch just because they happened to be a good hit in docking. In practice, docking is thus typically used either with libraries of compounds that can be bought, or compounds that can be synthesized cheaply by combining a few functional units. In this setting the computational cost of docking is not any limiting factor - and once we have a hit the work is rather focused on refining it (because we don't need the best possible hit in universe - just something that works).
@@eriklindahl I don't disagree with what you say. It's just that in the video, your justification for moving to "higher accuracy" FEP calculations is because "you ran out" of molecules to compute using docking. I have a problem with this rationale.
I feel the video would have been more accurate if instead you said, that we now move to a higher theory, first principle approach, which is limited in scope but more accurate than docking.