Non-linear interactions among single nucleotide polymorphisms (SNPs), genes, and pathways play an important role in human diseases, but identifying these interactions is a challenging task. Neural networks are state-of-the-art predictors in many domains due to their ability to analyze big data and model complex patterns, including non-linear interactions. In genetics, visible neural networks are popular as they provide insight into the most important SNPs, genes, and pathways for prediction. Visible neural networks use prior knowledge (e.g., gene and pathway annotations) to define node connections in the network, making them sparse and interpretable. Currently, most of these networks provide measures for the importance of SNPs, genes, and pathways but do not provide information about interactions. In this paper, we explore different methods to detect non-linear interactions with visible neural networks. We adapt and speed up existing methods, create a comprehensive benchmark with simulated data from GAMETES and EpiGEN, and demonstrate that these methods can extract multiple types of interactions from trained neural networks. Finally, we apply these methods to a genome-wide case-control study of inflammatory bowel disease and find high consistency of the epistasis pairs candidates between interpretation methods. The follow-up association test on these candidates identifies seven significant epistasis pairs.

Methane hydrate dissociation kinetics can be inhibited in NaCl solutions; however, this effect is reversed by promoting bubble formation that enhances dissociation. The negative and positive effects of inorganic salt injection on gas production from hydrate-bearing sediments are still controversial. Here, molecular dynamics simulations were performed to investigate the characteristics of NaCl solution invasion into hydrate-occupied nanopores and the effects on the confined hydrate dissociation kinetics. Two initial configurations comprising liquid and silica pore phases were studied with a low or high NaCl concentration, respectively. The results show that, under the simulation conditions, salt invasion decelerated hydrate dissociation within the silica pore as NaCl invasion into the pore is stepwise. Initially, few ions can diffuse into the pore phase, and gas nanobubbles form on the solid surface mainly via confinement and surface effects, independent of NaCl solution invasion. Subsequently, gradual salt diffusion immersed the residual hydrate in the salt solution and hindered hydrate decomposition until the dissociation finished. More ions could diffuse into the pore phase at the high NaCl concentrations with a low diffusion efficiency, leading to surface nanobubble growth toward the residual hydrate and somewhat accelerated hydrate dissociation. This severely hinders the escape of released methane from the pore. This study yields molecular-level insight into the origin of the negative effect of salt invasion on hydrate dissociation, which should be avoided during gas production from hydrate reservoirs with low permeabilities via salt injection combined with thermal stimulation.