Inverse resolution of spatially varying diffusion coefficient using Physics-Informed neural networks.

Resolving the diffusion coefficient is a key element in many biological and engineering systems, including pharmacological drug transport and fluid mechanics analyses. Additionally, these systems often have spatial variation in the diffusion coefficient which must be determined, such as for injectable drug-eluting implants into heterogeneous tissues. Unfortunately, obtaining the diffusion coefficient from images in such cases is an inverse problem with only discrete data points. The development of a robust method that can work with such noisy and ill-posed datasets to accurately determine spatially-varying diffusion coefficients is of great value across a large range of disciplines. Here, we developed an inverse solver that uses physics informed neural networks (PINNs) to calculate spatially-varying diffusion coefficients from numerical and experimental image data in varying biological and engineering applications. The residual of the transient diffusion equation for a concentration field is minimized to find the diffusion coefficient. The robustness of the method as an inverse solver was tested using both numerical and experimental datasets. The predictions show good agreement with both the numerical and experimental benchmarks; an error of less than 6.31% was obtained against all numerical benchmarks, while the diffusion coefficient calculated in experimental datasets matches the appropriate ranges of other reported literature values. Our work demonstrates the potential of using PINNs to resolve spatially-varying diffusion coefficients, which may aid a wide-range of applications, such as enabling better-designed drug-eluting implants for regenerative medicine or oncology fields.

Interfacial Dynamics in Dual Channels: Inspired by Cuttlebone.

The cuttlebone, a chambered gas-filled structure found in cuttlefish, serves a crucial role in buoyancy control for the animal. This study investigates the motion of liquid-gas interfaces within cuttlebone-inspired artificial channels. The cuttlebone's unique microstructure, characterized by chambers divided by vertical pillars, exhibits interesting fluid dynamics at small scales while pumping water in and out. Various channels were fabricated with distinct geometries, mimicking cuttlebone features, and subjected to different pressure drops. The behavior of the liquid-gas interface was explored, revealing that channels with pronounced waviness facilitated more non-uniform air-water interfaces. Here, Lyapunov exponents were employed to characterize interface separation, and they indicated more differential motions with increased pressure drops. Channels with greater waviness and amplitude exhibited higher Lyapunov exponents, while straighter channels exhibited slower separation. This is potentially aligned with cuttlefish's natural adaptation to efficient water transport near the membrane, where more straight channels are observed in real cuttlebone.

Characterizing viral samples using machine learning for Raman and absorption spectroscopy.

Machine learning methods can be used as robust techniques to provide invaluable information for analyzing biological samples in pharmaceutical industries, such as predicting the concentration of viral particles of interest in biological samples. Here, we utilized both convolutional neural networks (CNNs) and random forests (RFs) to predict the concentration of the samples containing measles, mumps, rubella, and varicella-zoster viruses (ProQuad®) based on Raman and absorption spectroscopy. We prepared Raman and absorption spectra data sets with known concentration values, then used the Raman and absorption signals individually and together to train RFs and CNNs. We demonstrated that both RFs and CNNs can make predictions with R2 values as high as 95%. We proposed two different networks to jointly use the Raman and absorption spectra, where our results demonstrated that concatenating the Raman and absorption data increases the prediction accuracy compared to using either Raman or absorption spectrum alone. Additionally, we further verified the advantage of using joint Raman-absorption with principal component analysis. Furthermore, our method can be extended to characterize properties other than concentration, such as the type of viral particles.

How a raindrop gets shattered on biological surfaces.

Many biological surfaces of animals and plants (e.g., bird feathers, insect wings, plant leaves, etc.) are superhydrophobic with rough surfaces at different length scales. Previous studies have focused on a simple drop-bouncing behavior on biological surfaces with low-speed impacts. However, we observed that an impacting drop at high speeds exhibits more complicated dynamics with unexpected shock-like patterns: Hundreds of shock-like waves are formed on the spreading drop, and the drop is then abruptly fragmented along with multiple nucleating holes. Such drop dynamics result in the rapid retraction of the spreading drop and thereby a more than twofold decrease in contact time. Our results may shed light on potential biological advantages of hypothermia risk reduction for endothermic animals and spore spreading enhancement for fungi via wave-induced drop fragmentation.

Efficient co-cultivation of human fibroblast cells (HFCs) and adipose-derived stem cells (ADSs) on gelatin/PLCL nanofiber.

In this study, we investigated whether the nanofibers produced by natural-synthetic polymers can probably promote the proliferation of co-cultured adipose-derived stem cells/human fibroblast cells (ADSs/HFCs) and synthesis of collagen. Nanofiber was fabricated by blending gelatin and poly (L-lactide co-ɛ-caprolactone) (PLCL) polymer nanofiber (Gel/PLCL). Cell morphology and the interaction between cells and Gel/PLCL nanofiber were evaluated by FESEM and fluorescent microscopy. MTS assay and quantitative real-time polymerase chain reaction were applied to assess the proliferation of co-cultured ADSs/HFCs and the collagen type I and III synthesis, respectively. The concentrations of two cytokines including fibroblast growth factor-basic and transforming growth factor-ß1 were also measured in culture medium of co-cultured ADSs/HDCs using enzyme-linked immunosorbent assay assay. Actually, nanofibers exhibited proper structural properties in terms of stability in cell proliferation and toxicity analysis processes. Gel/PLCL nanofiber promoted the growth and the adhesion of HFCs. Our results showed in contact co-culture of ADSs/HFCs on the Gel/PLCL nanofiber increased cellular adhesion and proliferation synergistically compared to non-coated plate. Also, synthesis of collagen and cytokines secretion of co-cultured ADSs/HFCs on Gel/PLCL scaffolds is significantly higher than non-coated plates. To conclude, the results suggest that Gel/PLCL nanofiber can imitate physiological characteristics in vivo and enhance the efficacy of co-cultured ADSs/HFCs in wound healing process.

Bending, twisting and flapping leaf upon raindrop impact.

Dynamics of drop impact on soft surfaces has drawn a lot of attention for its applications and is motivated by natural examples like raindrop impact on a leaf. Previous studies have focused on categorizing the bending motion observed, using cantilever beam theory, but the complex dynamic response shown by a leaf involving other degrees of motions like torsion about the petiole, remains yet to be understood. In this study, we demonstrated that the complex response of a superhydrophobic Katsura leaf upon raindrop impact can be decomposed into simple single degree-of-freedom linear modes of bending and torsion, modeled as damped harmonic oscillators. Our theoretical estimates were in good agreement with experimental measurements of the frequency and maximum amplitude of bending and torsional modes. We also illustrated the energy transfer from the raindrop to these modes as a function of the impact location, which may shed light on the design of potential raindrop energy harvesting devices mimicking a leaf's structure. Finally, we concluded with a brief description of an unresolved mode (i.e. flapping) and the limitations of our approach.

Association of HLA-DQB1 allelic sequence variation with susceptibility to systemic lupus erythematosus.

Systemic lupus erythematosus (SLE) is an autoimmune disease in which polymorphisms within the human leukocyte antigen (HLA) region have been associated to its etiology. We conducted this study to compare the HLA-DQB1 allelic sequence variation among SLE patients and controls in the northeast of Iran. Genomic DNA of 40 SLE patients and 83 healthy controls were amplified by Polymerase Chain Reaction with Sequence-Specific Primers technique (PCR-SSP). Seven serological subclasses of the HLA DQB1 were detected. Allele distribution comparison showed in the SLE group a significant increase of HLA DQ6 (*0601-*0609) (p=0.006); whereas alleles HLA DQ7 (*0301-*0304) were significantly decreased (p=0.005). Combination of DQ5 (*0501-*0504)-DQ6 (*0601-*0609) was increased in patients. These results suggest that DQ6 is the dominant HLA DQB1 allele probably associated with genetic susceptibility to SLE in the northeast of Iran. The association supports the importance of ethnic background and indicates the importance of various genes that has been observed in different SLE populations.