Self-assembly (SA) may be defined as the spontaneous organization of nanoparticles (NPs) into ordered structures by non-covalent interactions. The interactions responsible for SA are strictly local, i.e., the nanostructure builds itself from the NPs themselves. Generally, the SA interactions are thought weak comprising van der Waals, capillary, and hydrogen bonds that are about an order of magnitude less energetic than covalent bonds. Moreover, weak interactions are limited in to a typical range < 1 nm. In biological cells, SA occurs over a range of 1 - 10 microns, and therefore SA by weak interactions simply is not possible.
How then does SA at long range occur?
To answer this question, the SA of inanimate NPs into nanostructures may be used  to mimic that of proteins in biological systems. Inanimate SA suggests electrostatic interactions are the long-range mechanism in biological systems, specifically dipole-dipole interactions. Provided NP dipoles are activated, the dipole-dipole interactions between NPs lead to rapid SA.
One way of activating the dipoles uses an external source of EM radiation called light induced self-assembly (LISA). EM stands for electromagnetic. In Au NPs a photoactive dipole ADT coating has been shown  to assemble and dis-assemble SA depending on whether the ADT dipoles are activated by UV or VIS light. Absent external EM radiation, the SA of CdTe NPs  having their own dipoles are generally assumed activated by thermal fluctuations. Depending on the charge added to the NP, simulations show SA proceeds by forming sheets or chains. Initially, charge-dipole attraction governs SA, but short-range van der Waals attraction controls the final configuration.
Recently, experiments  show chicken egg white lysozyme (CEWL) under external UV radiation dissociates native disulfide bonds and mediate drastic conformational changes that expose the hydrophobic residues in partially unfolded molecules. Subsequently, the partially unfolded molecules under UV radiation undergo under rapid SA into globular protein aggregates.
Problem and Resolution
Immunogenicity of therapeutic drugs under UV radiation does not occur naturally except perhaps near the skin of the human body exposed to the UV content of sunlight. Protein aggregation based on UV radiation in the CEWL experiment  is therefore not relevant to clinical immunogenicity unless a source of UV radiation can be identified in the body.
What is the source of UV radiation in body fluids?
Immunogenicity is proposed mediated by UV created inside the protein aggregates themselves. The UV is a consequence of QM that requires the heat capacity of the atom to vanish in submicron aggregates. QM stands for quantum mechanics.
Lacking heat capacity by QM, the aggregate cannot increase in temperature to conserve the absorption of EM energy from the continuous collisions of water molecules in body fluids. Instead, conservation of absorbed EM energy proceeds by QED inducing the creation of EM radiation within the aggregate. QED stands for quantum electrodynamics. QED induced UV in protein aggregates finds similarity with natural and manufactured NPs that have been linked to DNA damage by the emission of low-level UV radiation See. http://www.prlog.org/
The MD simulation of protein aggregation by UV from the proteins themselves was simulated using the Verlet Algorithm under periodic boundaries. The initial configuration was taken as the FCC geometry of a crystal with 108 NPs. The NPs having diameter d = 2 nm were assumed to have protein density of 1200 kg/m3 and dispersed in the computation box of about 20 nm3 corresponding to a NP density of 100 kg/m3. Water molecules were excluded. The simulations were performed at atmospheric pressure and temperature assuming a typical protein polarizability of 1x10-30 m3. The force between protein NPs was mediated by the polarizability of the protein in the gradient of the energy density of the UV radiation field of all NPs. See Paper entitled: “Aggregation of Proteins by Polarization,”
The Visual Molecular Dynamics image of natural aggregation of the globular protein is shown in the thumbnail. Similar to the CEWL experiment , the SA once started is rapid in the formation of 10 nm globular proteins. The final configuration determined by short-range van der Waals and dipole-dipole interactions is unimportant for the purposes here.
Summary and Conclusions
1. SA in long-range attraction between inanimate or biological NPs is mediated by the polarizability of the NP in the gradient of the EM radiation field produced by the NPs themselves. Once the NPs are close to each other, SA proceeds by Van der Waals and dipole-dipole interactions.
2. Activation of NP dipoles in SA does not occur as commonly thought by thermal fluctuations because the NPs lack the heat capacity to allow the dipoles to oscillate with the temperature. Instead, dipole activation occurs from the EM radiation produced within the NPs themselves.
3. The EM radiation produced by the NPs is a consequence of QM and QED. Heat capacity of the NP is precluded by QM. EM energy from inelastic collision of water molecules absorbed by the NPs therefore cannot be conserved by an increase in temperature. Instead, QED conserves the absorbed EM energy by producing EM radiation. The source of SA is therefore inelastic collisions of water molecules.
4. Immunogenicity by QED induced UV radiation from the proteins themselves requires the refractive index of the protein to be greater than that of water. Hence, drugs that enhance absorption of water by proteins should reduce the tendency of immunogenicity.
 R. Klajn, et al., “Light-controlled self-assembly of reversible and irreversible nanoparticle suprastructures,’
 N. Kotov, “Inorganic Nanoparticles as Protein Mimics,” Science, Vol. 330, pp. 189 (2010).
 Z. Zhang, et al., “Simulations and Analysis of Self-Assembly of CdTe Nanoparticles into Wires and Sheets”, Nano Letters, Vol. 7, pp. 1670 (2007).
 J. Xie, et. al., “Mechanistic insight of photo-induced aggregation of chicken egg white lysozyme: The interplay between hydrophobic interactions and formation of intermolecular disulfide bonds,” Proteins, vol. 79, pp. 2505–2516, 2011.