ReaDDy Simulations for actin bend/twist coupling figure

[blairlyons]

Use Case

As an author for the actin bend/twist coupling micro-publication, I need data to complete the accompanying figure.

Acceptance Criteria

.simularium trajectories for replicates of all the parameter sets needed for the figure, and plots for pitch and bending for all the replicates of each set.

Details

figure mockup:

ReaDDy Model:

Model:

• start with straight linear filament with length ~= 150 nm
• pointed end is fixed (first 3 monomers to fix both position and rotation) and displaced
• barbed end position is fixed and rotation is either free or fixed (either 1 or 3 monomers fixed respectively)
• total simulation time is 10 us
• final magnitude of displacement is 42 nm, applied linearly over the time course of simulation
• displacement is either tangent (moving pointed end toward barbed end) or normal/arc (moving pointed end along arc normal to axis):
  
• in total 4 conditions:
  • fixed barbed rotation, tangent displacement of pointed
  • fixed barbed rotation, normal/arc displacement of pointed
  • free barbed rotation, tangent displacement of pointed
  • free barbed rotation, normal/arc displacement of pointed
• 10 replicates of each (?)

Metrics:

• end displacement distance (independent var)
• angles between normal vectors for each pair of monomers
• total twist: sum of monomer normal angles along length of filament
• bond lengths between each monomer, both laterally and longitudinally

Plots (per trajectory):

• total twist vs end displacement
• bond lengths (lat and long) vs end displacement (normalize bond lengths relative to theoretical lengths, plot average ± std)

Reference: https://www.sciencedirect.com/science/article/pii/S0006349510008532

[blairlyons]

@mattakamatsu does this specification for the ReaDDy models look right? Am I forgetting anything?

Actually, which end do we want to displace? If we displace an end, in the current implementation, at least its position is fixed

[mattakamatsu]

Ah right - okay so really we'll have two conditions (the ones you simulated) - move with free rotation or no rotation. I updated the figure mockup accordingly

[mattakamatsu]

Updated figure mockup showing the output metrics we are looking for:

Proxy for filament twist: angular displacement between subunits
Proxy for filament bend: displacement between subunits

• longitudinal
• lateral

[blairlyons]

@mattakamatsu (EDIT: I just updated the description above with) what I'm planning to implement, just want to triple check we're on the same page before I get started. lmk if that sounds ok!

[mattakamatsu]

@blairlyons That all looks accurate to me! I like the idea of normalizing bond length to starting or theoretical lengths, and plotting the average ± std at first. I think this should give us enough info to point our way toward a quantitative comparison with Schramm et al.

[blairlyons]

@mattakamatsu here's some initial results! 🎉 I'm not sure the twist metric is useful when there's a lot of bending?

• Displace radially, barbed rotation fixed
• Displace radially, barbed rotation free
• Displace tangentially, barbed rotation fixed
• Displace tangentially, barbed rotation free

If this looks ok, I can run the replicates in the cloud

[mattakamatsu]

@blairlyons This looks fantastic!! The trending upward of the twist metric is really interesting. I think we'll see more twist if we displace all the way to a 90˚ bend. Which I think is 173 nm? https://www.omnicalculator.com/math/arc-length
Otherwise I think it's ready to run replicates. I'm also ready to share with Enrique to help us compare quantitatively to De La Cruz 2010 and Schramm 2017, if that's ok with you!

[blairlyons]

ok I bent the radial ones 90˚, the displacement only gets to 105 nm, but I'm using a radius of 75 nm and it's the euclidean distance on that plot, not the arc length. I also added some vector projection to get only the component of the normal perpendicular to the tangent, thinking this would remove the bending angle from the twist metric. The two metrics diverge in the new runs, we should probably talk about how to calculate this metric. Maybe it's fine without the projection, but when I look at the spatial output it doesn't seem to match...

Those links are updated:

• Displace radially, barbed rotation fixed
• Displace radially, barbed rotation free
• Displace tangentially, barbed rotation fixed
• Displace tangentially, barbed rotation free

I'm fine with you sharing with Enrique whenever you think it will be useful, probably would be good to know that these are comparable before I run in the cloud. thanks!