Incredible work spearheaded by lead author Nir Nitskansky, and by Kessem Clein, who created the initial proof of concept for mMSMs. Full though not yet final version, now in Bioinformatics -
academic.oup.com/bioinformati...
Github with code and tutorial - github.com/ravehlab/mMSM
3/3
We also introduce mMSM-explore, an unsupervised algorithm for efficiently generating multiscale Markov state models (mMSMs) of biomolecular systems, operating through multiscale adaptive sampling with on-the-fly identification of temporally metastable states:
2/3
We introduce multiscale Markov state models (mMSMs), compact summaries of molecular dynamics (MD) simulation trajectories simultaneously capturing multiple temporal resolutions. In this example, we used mMSM to describe the folding process of the HP35 miniprotein:
1/3 #mMSM #MD
Explained here in more detail:
bsky.app/profile/rave...
Note in the above illustrative movie, the density and scale of the disordered FG repeats lining the NPC’s central channel are lowered for visual clarity. For a more realistic view, check out our recent integrative model of nucleo-cytoplasmic transport, now in PNAS
#NPC
www.pnas.org/doi/10.1073/...
Cool work by Rod Lim’s lab on transport through nuclear pore complexes, the gateways to the nucleus, with contributions from our own Roi Eliasian, who developed a simulator to directly compare imaging data to our modeling.
#npc #transport
www.nature.com/articles/s41...
m.youtube.com/watch?v=9FPX...
The integrative model enables us to visualize how the flexible FG chains lining the NPC’s central channel filter molecular traffic between the nucleus and the cytoplasm at picosecond resolution, well beyond the capabilities of any existing imaging technology.
#NPC #transport
The integrative model enables us to visualize how the flexible FG chains lining the NPC’s central channel filter molecular traffic between the nucleus and the cytoplasm at picosecond resolution, well beyond the capabilities of any existing imaging technology.
#NPC #transport
This was a true team effort across HUJI, UCSF/QBI, Rockefeller, Einstein, Basel, and UCSD/HHMI. Immense thanks to our colleagues and collaborators at the Sali, Rout, Cowburn, Villa and Lim labs.
bsky.app/profile/rave...
Aberrant nucleocytoplasmic transport is associated with disease, from cancer to neurodegeneration and viral infection. A mechanistic, quantitative model opens doors to diagnostics and therapeutic design, for both native pores and their artificial mimics. /
The basic transport mechanism is like buoyancy in water: inside the NPC, flexible FG chains create an entropic push that repels unescorted molecules from the mid-plane. NTRs add molecular ballast via FG binding, counteracting buoyancy and ferrying cargo through.
A key outcome: we identify 10 molecular design features that together yield speed, selectivity, and robustness for nucleocytoplasmic transport through the NPC, and predict how tuning them shifts performance.
The model recapitulates observed selectivities and fluxes across cargo sizes and receptor properties, and explains how massive cargo can still pass quickly when properly escorted.
The model recapitulates observed selectivities and fluxes across cargo sizes and receptor properties, and explains how massive cargo can still pass quickly when properly escorted.
Large cargoes are nonetheless granted passage through the pore if they carry the right “passport”: nuclear transport receptors (NTRs). NTRs make many fast, weak, transient handshakes with FG chains, using a slide-and-exchange mechanism that allows them to glide smoothly from chain to chain.
Our answer centers on an entropic barrier created by a dense, dynamic “forest” of flexible FG-repeat protein chains inside the NPC’s central channel. Large molecular cargoes punch holes in this dynamic thicket, incurring an energetic penalty for free passage through the pore.
Years of work, finally out! Nuclear Pore Complexes (NPCs) are the gateways to the nucleus, but how can they be both rapid and picky, even for very large cargoes? To answer this question, we built the most detailed data-driven model to date of transport through NPCs. /
www.pnas.org/doi/10.1073/...
Check out the TEMPO algorithm from our lab by the incredible Reshef Mintz. By recursively forecasting future timesteps from recent ones across scales, TEMPO accelerates the costly integration step of molecular dynamics, even for complex biomolecular processes such as nucleocytoplasmic transport.
Kudos to #kaplanlab@huji for this exciting work, identifying 1000s of genomic regions showing allele-specific DNA methylation, including novel cell-type-specific imprinted regions.
Characterizing single-cell and spatial data structure and identifying gaps in our knowledge of their annotations by analyzing deep learning training dynamics!
Led by @jonathankarin.bsky.social and Reshef Mintz, with
@ravehlab.bsky.social
1/3 Annotatability is out in Nature Computational Science! With Reshef Mintz,
@ravehlab.bsky.social
and @mornitzan.bsky.social
www.nature.com/articles/s43...
