Folding@home Sprint 5 focuses on compounds that build on the benzopyran-isoquinoline lead MAT-POS-b3e365b9-1 and MAT-POS-8a69d52e-7 to engage the P1’ pocket.
See the Folding@home blog post for more information.
Hi John, how predictive were the free energy calculations of Δ(pIC50) values when used prospectively (i.e. in previous sprints)?
Peter,
We’re working on adding a “Retrospective” tab to each of the Sprint dashboards—Sprint 5 will come up first. But here’s a quick figure on how the top compounds from Sprint 1 did:
We’ve made a number of improvements (and fixed a lot of bugs!) since then.
New sprints from Sprint 6 on should first show retrospective performance on existing compounds and then predicted performance on new compounds automatically.
Thanks, John, and it might be an idea to report pIC50 (rather than IC50) since this makes it easier to quickly perceive differences between predicted and measured potency at a glance. In case you’ve not already seen it, this comment, in which minimum significant ratio (MSR is discussed, may be of interest.
Hi John @JohnChodera
I’d like to highlight a potential opportunity (which the design team may have already considered) for application of FEP calculations and I’ll mention @edgriffen @mc-robinson @RGlen @frankvondelft @alphalee @londonir in case this is of interest.
A number of designs (some have already been synthesized) feature a ring that connects the amidic nitrogen of the 3-aminopyridine-like substructure with the benzylic carbon. The structural variation for this ‘cyclized’ scaffold design theme is within the core of the structure rather than at the periphery (makes the synthesis more difficult) while the perturbations are typically small in terms of molecular size (FEP may perform better).
The lactams PET-UNK-c9c1e0d8-3 and PET-UNK-c9c1e0d8-4 are equipotent to the parent compound ADA-UCB-6c2cb422-1. Both would provide access to the S1’ subsite although PET-UNK-c9c1e0d8-3 was specifically designed to for targeting the catalytic cysteine (see PET-UNK-3e354a91-1).
The cyclic imide VLA-UCB-1dbca3b4-10 and the dihydrouracil VLA-UCB-1dbca3b4-18 each has an additional carbonyl group in the ring. This makes them hybrids of the 3-aminopyridine-like and quinolone series as discussed previously. I would anticipate that the potency advantage of isoquinoline over pyridine will be diminished when the heteroaromatic ring is flanked by two carbonyl groups.
The chiral center in the ‘cyclized’ scaffolds is a concern because of its potential for racemization (it is adjacent to a carbonyl group) and locking the configuration may result in loss of potency. The hydantoin VLA-UCB-29506327-1 has actually been synthesized although it is not especially potent (it is equipotent to PET-UNK-c9c1e0d8-4). I would anticipate non-additivity in SAR (e.g. potency advantage of the dihydrobenzopyran might be diminished for the spiro ring fusion). My view is that hydantoins such as VLA-UCB-29506327-1 are likely to be less potent than dihydrouracils such as VLA-UCB-34f3ed0c-12. One way to address the problems associated with the chiral center is to simply eliminate it (see designs VLA-UCB-1dbca3b4-13 | PET-UNK-e8933450-1 | PET-UNK-4dc48bbe-4 ).
I hope this gives a feel for ‘cyclized’ scaffold structural variation that might be explored using FEP. If this is of interest, I’ll be happy to discuss further and to provide input to the selection of structures for evaluation.