Search

Background: DNA double-strand breaks (DSBs) are among the most deleterious DNA lesions, and they can cause cancer if improperly repaired. Recent chromosome conformation capture techniques, such as Hi-C, have enabled the identification of relationships between the 3D chromatin structure and DSBs, but little is known about how to explain these relationships, especially from global contact maps, or their contributions to DSB formation. Results: Here, we propose a framework that integrates graph neural network (GNN) to unravel the relationship between 3D chromatin structure and DSBs using an advanced interpretable technique GNNExplainer. We identify a new chromatin structural unit named the DNA fragility–associated chromatin interaction network (FaCIN). FaCIN is a bottleneck-like structure, and it helps to reveal a universal form of how the fragility of a piece of DNA might be affected by the whole genome through chromatin interactions. Moreover, we demonstrate that neck interactions in FaCIN can serve as chromatin structural determinants of DSB formation. Conclusions: Our study provides a more systematic and refined view enabling a better understanding of the mechanisms of DSB formation under the context of the 3D genome.

Due to the high energy consumption in CO2capture by amine absorbents, CO2phase change absorbent, CPCA, has attracted a great attention in research institutes recently. Two phases are formed after CO2absorption with CPCA as the absorbent, and only the CO2-rich phase goes to the stripper, which could potentially reduce the energy consumption. In this work, the CPCAs composed of MEA/1-propanol/H2O were developed based on the principles of aqueous two-phase systems. The effects of temperature, amine concentration, and total CO2loading on the phase separation of CPCAs were studied. CO2loading and 1-propanol distributions in two phases under different experimental conditions were measured. The results indicated that the total CO2loading is the key factor affecting the distribution of components in two phases. Moreover, in comparison to traditional 30 wt % MEA, the initial absorption rates of all CPCAs were higher than that of MEA, and its maximum value was 5 times higher. The cyclic capacity of the CPCA with 30% MEA + 30% 1-propanol + 40% H2O was 1.70 mol/kg, which was 52% increased comparing to traditional 30 wt % MEA.

Ovalbumin (OVA)/sodium carboxymethylcellulose (CMC) colloidal particles were prepared with different compactness and morphologies by regulating the interaction between proteins and polysaccharides during heating. Electrostatic interactions between the amine groups of OVA (−NH3+) and carboxyl groups of CMC (−COO–) enhanced complex formation. The protein conformation change benefited the hydrophobic interaction between the particles. Proteins in colloidal particles were unfolded/folded under thermal induction to form aggregates having more β-sheet structures. When the OVA/CMC ratio was 1:2, the initially loosely connected OVA/CMC aggregation changed into a uniform sphere between 25 and 90 °C. The mass ratio of OVA to CMC within the final colloidal particle (90 °C) was about 1:1.4. The OVA/CMC particle stability was maintained with hydrogen bonding, hydrophobicity, and disulfide bond. When OVA levels were predominant, OVA and CMC developed an approximately hollow sphere. Moreover, the final colloidal particle composition showed the OVA-to-CMC ratio as 3:1 (w/w). OVA bound into colloidal particle pores to increase compactness. Moreover, OVA and CMC bound to the colloidal particle while the particle shrank, thereby increasing the compactness of colloidal particles. There was a significant decrease in ABTS•+scavenging activity of curcumin compared with that of the particles with a ratio of 1:2. Thus, the rational adjustment of the structure of colloidal particles could effectively enhance their functional characteristics, providing a new way for the controlled release of the active ingredients.

Positive and unlabeled (PU) learning is the problem in which training data contains only PU samples. Although PU learning is widely used in real-world applications, its model selection remains challenging. Specifically, traditional model selection methods are often highly sensitive to the class prior as well as the data size, resulting in human overhead in hyperparameter optimization. In this paper, we present a method called ODE (robust model selection) for robust model selection in PU learning. Two novel model evaluators based on the integral probability metric and area under the curve, which are free of the class prior, are introduced, and a variance reduction method is further employed to improve the quality of model selection. In addition, we perform model selection under user-defined constraints and propose a fast halving-style searching algorithm to efficiently identify the most promising model configuration. Extensive empirical studies demonstrate that our proposed method performs more robustly and is more computationally efficient than many state-of-the-art methods.