Results of the numerical simulations.

Typical pressure distribution in case of free-slip (a) and no-slip (b) boundary conditions imposed on the bottom of the Petri dish. The flow field has an axial symmetry, and only the right half of the geometry is shown in the side views. The real-life velocity profile (and other integral quantities such as the pressure drop) is expected to lie between these two extreme cases. Distance between the tip of the micropipette and the bottom of the Petri dish: H = 10 µm. Flow rate: 6 µl/s. To validate the results of simulations we compared the simulated flow rate of the micropipette to the experimental values as a function of the vacuum value with H = 5 µm (c) and H = 10 µm (d) taking into consideration corrections due to gravity, pressure drop in the PTFE tube and the flow velocity in the micropipette (Figure S1 in File S1). Simulation with free-slip condition on the bottom of the Petri dish proved to be a better approximation of the experiments than the no-slip simulations. Thus we determined the lifting force (e) acting on the hemisphere model of the cell on the basis of the free-slip simulations as a function of the vacuum applied to the micropipette. With a linear fitting we found the following relation between the hydrodynamic lifting force (FL) and the vacuum (V) applied to the micropipette: FL = 0.172 [nN/Pa] * V +311 [nN] (R2 = 0.996) if H = 5 µm. FL = 0.071 [nN/Pa] * V +961 [nN] (R2 = 0.999) if H = 10 µm. We used these coefficients to convert the experimental vacuum values to an estimated lifting force.