Hi all,

Relative noobie here, looking for general advice on how I might approach modeling a cancer cell that looks like this:
https://www.mskcc.org/sites/default/files/styles/width_800/public/node/167374/3x2/sciencesource_ss21694540_1200x800.jpg [www.mskcc.org]
or this:
https://static.independent.co.uk/s3fs-public/thumbnails/image/2016/02/15/21/1-t-cell-corbis.jpg?w968h681 [static.independent.co.uk]
One physical feature of the cell that I’d like to pull off convincingly is the sticky-looking tendrils that shoot out from the membrane, that look a bit like viscous liquid has been shot out from the surface of the cell.

Though something like the image in the above links would be ideal, I would settle for something like this:
https://lymphomanewstoday.com/wp-content/uploads/2018/09/shutterstock_565670083_zps0eglo9ri-1000x480.jpg [lymphomanewstoday.com]
I would want to avoid something like this (where the tendrils resemble veins, or maybe octopus tentacles:
https://www.segnidalcielo.it/wp-content/uploads/2017/08/GoldNanoparticles.jpg [www.segnidalcielo.it]

Since for now I only actually need a static model, my first inclination would be to start with a simple POP setup, have some particles explode off a spherical surface and append a trail SOP to get the tendril trails. I’d run the sim and use a single frame from it as the cancer cells base geo. This only gets me so far though, as there are other important visual details that I’ll refrain from talking about until I some experienced person can tell me that this is definitely the way to go.

My other inclination, which I realize might be inviting insanity, is to create a “fluid explosion”, using a FLIP setup with some velocity noise coming from the SOP level, and playing with DOP node settings to get the right kind of viscous/stringy effect from the particles. From there, again, I’d run the sim and use a single frame of geo for the base for my model.
I’ve so far made an attempt at this (file attached), though have only tried seeing what I could get by using nothing but some noised-up velocity on the SOP source, some DOP gravity and tweaking settings in DOP Source Volume and Solver nodes, pertaining particularly to viscosity and surface tension (though I’m not sure the latter is relevant). I’m able to get some neat looking effects, but not really the ones I’m specifically after.
What I’d wonder here is whether
this ‘explosion of viscous fluid with a very high tensile strength’ could even be achieved with a fluid sim, and if so, would I need to push/pull the “sticky/stringy” fluid outward by having the FLIP particles stick to a set of “collision” particles which are then launched outward, pulling some of the fluid along with it.
Or, could this effect be achieved by simply applying velocity to a Fluid Source SOP with very low particle separation, and then relying on settings available in DOPs nodes to get the long / stretchy tendrils.

In any event, I’m certainly open to suggestions even if they’re completely different approaches to what I’ve described above. Thanks ahead of time to anyone responding!


Nik pF

I think that this could be easily modeled using a few different noises for the main shape. Maybe I would use an anti-aliased noise to get the large displacements and then I would use the unified noise for the fine displacement, paying attention to the gradient and folding section in a VOP on a very high res sphere or you could also use volumes and do this all in a Volume Vop if you needed that. For the strands, I would look at the gnarly or branch example in an L-system Sop and modify that to get an approximation of that style of branching. I would then use a for loop to instance the Lsystem to scattered points with “iteration” on a few properties to get variety, with some points having a random, controllable offset from the surface. There ae all kinds of ways of surfacing to get your desired result, maybe the easiest would be a polywire with some noise bound to a “width” attribute that I would then use with a subdivide to smooth.

I would stay away from fluids but you can always scatter points on the strand wires, maybe use a pointRecplicate on some points and then use a particleFluidSurface but I think you would get mixed results that might not easily be controllable.

Jordan