HelixADMET: a robust and endpoint extensible ADMET system incorporating self-supervised knowledge transfer.

MOTIVATION: Accurate ADMET (an abbreviation for 'absorption, distribution, metabolism, excretion and toxicity') predictions can efficiently screen out undesirable drug candidates in the early stage of drug discovery. In recent years, multiple comprehensive ADMET systems that adopt advanced machine learning models have been developed, providing services to estimate multiple endpoints. However, those ADMET systems usually suffer from weak extrapolation ability. First, due to the lack of labelled data for each endpoint, typical machine learning models perform frail for the molecules with unobserved scaffolds. Second, most systems only provide fixed built-in endpoints and cannot be customized to satisfy various research requirements. To this end, we develop a robust and endpoint extensible ADMET system, HelixADMET (H-ADMET). H-ADMET incorporates the concept of self-supervised learning to produce a robust pre-trained model. The model is then fine-tuned with a multi-task and multi-stage framework to transfer knowledge between ADMET endpoints, auxiliary tasks and self-supervised tasks. RESULTS: Our results demonstrate that H-ADMET achieves an overall improvement of 4%, compared with existing ADMET systems on comparable endpoints. Additionally, the pre-trained model provided by H-ADMET can be fine-tuned to generate new and customized ADMET endpoints, meeting various demands of drug research and development requirements. AVAILABILITY AND IMPLEMENTATION: H-ADMET is freely accessible at https://paddlehelix.baidu.com/app/drug/admet/train. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Full Geometric Control of Hidden Color Information in Diffraction Gratings under Angled White Light Illumination.

Under white light illumination, gratings produce an angular distribution of wavelengths dependent on the diffraction order and geometric parameters. However, previous studies of gratings are limited to at least one geometric parameter (height, periodicity, orientation, angle of incidence) kept constant. Here, we vary all geometric parameters in the gratings using a versatile nanofabrication technique, two-photon polymerization lithography, to encode hidden color information through two design approaches. The first approach hides color information by decoupling the effects of grating height and periodicity under normal and oblique incidence. The second approach hides multiple sets of color information by arranging gratings in sectors around semicircular pixels. Different images are revealed with negligible crosstalk under oblique incidence and varying sample rotation angles. Our analysis shows that an angular separation of ≥10° between adjacent sectors is required to suppress crosstalk. This work has potential applications in information storage and security watermarks.

Handheld robotic needle holder training: slower but better.

BACKGROUND: Handheld robotic laparoscopic instruments fill the gap between robotic and conventional instruments, combining the advantages of degrees of freedom and low price. The difficulty and value in learning these new instruments require detailed investigation. METHODS: Forty novice surgeons with no laparoscopic experience were randomly assigned to two groups: conventional instrument group (Group Conv) and robotic instrument group (Group Rob). The same training protocol was used in both groups: after viewing a standard operation film, laparoscopic suture training was administered using the corresponding instruments. After each training period, surgeons were tested using a force-sensing test platform. Maximum force (MF) and impulse (IMP) of operators through each ring were recorded. Learning curves based on MF and IMP for both instruments were compared. Institutional review board approval is not needed for this study. RESULTS: MF and IMP of both groups decreased with increased training time; the learning curve of Group Conv decreased faster than that of Group Rob. When training time reached 13 h, the MF of Group Rob was significantly lower than that of Group Conv (P < 0.05), while IMP showed no significant difference between the two groups. CONCLUSIONS: Effective training reduces operator MF and IMP, possibly decreasing damage to tissues with both conventional and handheld robotic needle holders. Group Rob took longer to reach a plateau, but subsequently had lower suture tension than did Group Conv. MF is more sensitive than IMP for measuring performance progress.

Long noncoding RNA FLVCR1-AS1 aggravates biological behaviors of glioma cells via targeting miR-4731-5p/E2F2 axis.

Long noncoding RNAs (lncRNAs) display essential roles in cancer progression. FLVCR1-AS1 is a rarely investigated lncRNAs involved in various human cancers, such as hepatocellular carcinoma and lung cancer. However, its function in glioma has not been clarified. In our study, we found that FLVCR1-AS1 was highly expressed in glioma tissues and cell lines. And upregulation of FLVCR1-AS1 predicted poor prognosis in patients with glioma. Moreover, FLVCR1-AS1 knockdown inhibited proliferation, migration and invasion of glioma cells. Through bioinformatics analysis, we identified that FLVCR1-AS1 was a sponge for miR-4731-5p to upregulate E2F2 expression. Moreover, rescue assays indicated that FLVCR1-AS1 modulated E2F2 expression to participate in glioma progression. Altogether, our research demonstrates that the FLVCR1-AS1/miR-4731-5p/E2F2 axis is a novel signaling in glioma and may be a potential target for tumor therapy.

Identification of a novel series of azabenzimidazole-derived inhibitors of spleen tyrosine kinase.

Spleen Tyrosine Kinase (SYK) is a well-studied enzyme with therapeutic applications in oncology and autoimmune diseases. We identified an azabenzimidazole (ABI) series of SYK inhibitors by mining activity data of 86,000 compounds from legacy biochemical assays with SYK and other homologous kinases as target enzymes. A structure-based design and hybridization approach was then used to improve the potency and kinase selectivity of the hits. Lead compound 23 from this novel ABI series has a SYK IC50 = 0.21 nM in a biochemical assay and inhibits growth of SUDHL-4 cells at a GI50 = 210 nM.

Handheld laparoscopic robotized instrument: progress or challenge?

BACKGROUND: Handheld laparoscopic robotized instruments have been developed to combine the advantages of a robotic operation system and conventional laparoscopic instruments. Direct objective standards are needed to quantify surgeons' skill level and validate the advantages of new instruments. This study describes and objectively evaluates the use of a robotized instrument using a force-sensing test platform. METHODS: The test platform consists of 12 rings on a hypersensitive force sensor. Forty volunteers were recruited: the expert group included 20 laparoscopic experts and the novice group included 20 medical students in their 4th year without laparoscopic skills. The baseline of the two groups was identified using a conventional needle holder. Participants then repeated the test with the robotized needle holder after training. The maximum force and impulse of each ring were analyzed for each group. Institutional review board approval is not needed for this study. RESULTS: Significantly lower maximum force and impulse were observed in the expert group than in the novice group during the baseline test (all P < 0.05). After training, a significant difference remained in maximum force and impulse between the two groups using the robotized needle holder (all P < 0.05). Within each group, there was no difference in maximum force or impulse using the robotized needle holder after training compared to that on using the conventional needle holder (all P > 0.05). CONCLUSIONS: The maximum force and impulse recorded by a test platform can accurately identify participants' laparoscopic skill level. Six hours' training can ensure that experts master the use of the robotized needle holder, but this training session is too short for novices to improve their performance with a new instrument. The force-sensing test platform can reflect the suturing characteristics of users based on the skill level and is useful for laparoscopic suture training.

Suppression of NLRP3 inflammasome attenuates stress-induced depression-like behavior in NLGN3-deficient mice.

Depression, regulated by central nervous system (CNS), is a significant inflammatory disorder. Neuroligin3 (NLGN3) has been implicated in brain functions. In the study, a chronic unpredictable mild stress (CUMS) model in wild type (WT) or NLGN3-knockout (KO) mice was established to explore the role of NLGN3 in regulating depression and to reveal the underlying molecular mechanism. The results indicated that NLGN3-knockout markedly reversed the loss of body weight, the reduction of sucrose consumption, the decrease of immobile time in the forced swimming tests (FST) and tail suspension tests (TST) induced by CUMS paradigm. CUMS up-regulated corticosterone (CORT) in serum, and down-regulated serotonin (5-HT), norepinephrine (NE) and brain-derived neurotrophic factor (BDNF) in hippocampus of mice, which were significantly reversed by NLGN3 deficiency. The results further demonstrated that NLGN3-knockout improved the degenerative neurons in cortex and hippocampus of CUMS-treated mice, accompanied with a significant decrease of ionized calciumbinding adapter molecule 1 (Iba-1) and glial fibrillary acidic protein (GFAP) expressions. Additionally, NLGN3-KO mice challenged with CUMS showed a significant reduction of pro-inflammatory cytokines and chemokine, including tumor necrosis factor-alpha (TNF-α), interleukin-18 (IL-18), interleukin-1 beta (IL-1ß), interleukin-4 (IL-4), CC-chemokine ligand-1 (CCL-1) and CXC-chemokine ligand-1 (CXCL-1), in cortex, hippocampus and amygdala tissue samples. Western blot analysis suggested that NLGN3-knockout inhibited the activation of nod-like receptor protein 3 (NLRP3) inflammasome and its adaptor of apoptosis-associated speck like protein (ASC), and reduced the expression of Caspase-1, along with the inactivation of nuclear factor-κB (NF-κB) in CUMS-challenged mice. The role of NLGN3 in regulating depression in mice was confirmed in vitro using astrocytes stimulated by LPS that NLGN3 knockdown reduced LPS-induced inflammation. Importantly, the suppressive effects of NLGN3-knockdown on inflammatory response were reversed by NLRP3 or ASC over-expression in AST exposed to LPS. In sum, our findings indicated that suppressing NLGN3 played a potential antidepressant role in CUMS animal model by inactivating NLRP3 inflammasome, providing a new therapeutic avenue for depression.

Design and Evaluation of FBG-Based Tension Sensor in Laparoscope Surgical Robots.

Due to the narrow space and a harsh chemical environment in the sterilization processes for the end-effector of surgical robots, it is difficult to install and integrate suitable sensors for the purpose of effective and precise force control. This paper presents an innovative tension sensor for estimation of grasping force in our laparoscope surgical robot. The proposed sensor measures the tension of cable using fiber gratings (FBGs) which are pasted in the grooves on the inclined cantilevers of the sensor. By exploiting the stain measurement characteristics of FBGs, the small deformation of the inclined cantilevers caused by the cable tension can be measured. The working principle and the sensor model are analyzed. Based on the sensor model, the dimensions of the sensor are designed and optimized. A dedicated experimental setup is established to calibrate and test the sensor. The results of experiments for estimation the grasping force validate the sensor.

Variable Admittance Control Based on Fuzzy Reinforcement Learning for Minimally Invasive Surgery Manipulator.

In order to get natural and intuitive physical interaction in the pose adjustment of the minimally invasive surgery manipulator, a hybrid variable admittance model based on Fuzzy Sarsa(λ)-learning is proposed in this paper. The proposed model provides continuous variable virtual damping to the admittance controller to respond to human intentions, and it effectively enhances the comfort level during the task execution by modifying the generated virtual damping dynamically. A fuzzy partition defined over the state space is used to capture the characteristics of the operator in physical human-robot interaction. For the purpose of maximizing the performance index in the long run, according to the identification of the current state input, the virtual damping compensations are determined by a trained strategy which can be learned through the experience generated from interaction with humans, and the influence caused by humans and the changing dynamics in the robot are also considered in the learning process. To evaluate the performance of the proposed model, some comparative experiments in joint space are conducted on our experimental minimally invasive surgical manipulator.

[Research and Application of the ICP-MS Detection Technology for the Content of Nb and Ta in Geochemical Sample].

In order to provide the test analysis technology to support the exploration and development of niobium and tantalum resource, based on the special chemical properties of Nb and Ta in geochemical sample, we studied the detection methods for the content of Nb and Ta in geochemical sample by using inductively coupled plasma mass spectrometry (ICP-MS). The results show that the sample dissolution and instrumental parameter of ICP-MS, especially the former have significant influence? on detection results. Therefore, optimizing important parameters of sample dissolution is the key of the detection technology. The optimal parameters are that the weight of sample is 50 mg; the dosage of HF acid is 15 mL; the concentration of nitric acid and tartaric acid in the sample solution is 2% and 1.5%, respectively; the validity period of detection for sample solution ≤1 d. The detection method has been validated by the national geochemistry standard reference material. The precise and exaction of method meet the required of industry standards. The detection limits of method for Nb and Ta are 1.05 and 0.13 µg · g(-1), respectively. The experiment proved that the ICP-MS detection methods, which using certain preparation process of sample solutions, is suitable for accurate and rapid determination of Nb and Ta in geochemical sample, especially geochemical survey samples which with a large amount and low content of Nb and Ta.

Novel Human Milk Fat Substitutes Based on Medium- and Long-Chain Triacylglycerol Regulate Thermogenesis, Lipid Metabolism, and Gut Microbiota Diversity in C57BL/6J Mice.

Human milk is naturally rich in medium- and long-chain triacylglycerols (MLCT), accounting for approximately 30% of the total fat. However, infant formula fat is prepared using a physical blend of vegetable oils, which rarely contains MLCT, similar to human milk. The differences in MLCT between human milk and infant formulas may cause different lipid metabolisms and physiological effects on infants, which are unknown. This study aimed to analyze the metabolic characteristics of formula lipid containing novel human milk fat substitutes based on MLCT (FL-MLCT) and compare their effects with those of the physical blend of vegetable oils (FL-PB) on lipid metabolism and gut microbiota in mice. Compared with the FL-PB group, the FL-MLCT group showed increased energy expenditure, decreased serum triacylglycerol level, and significantly lower aspartate aminotransferase level, epididymal and perirenal fat weight, and adipocyte size. Moreover, the abundances of Firmicutes/Bacteroidota, Actinobacteriota, and Desulfovibrionaceae were significantly decreased in the FL-MLCT group. Novel human milk fat substitutes MLCT could inhibit visceral fat accumulation, improve liver function, and modulate the mice gut microbiota composition, which may contribute to controlling obesity.

Tunable VO2 cavity enables multispectral manipulation from visible to microwave frequencies.

Optical materials capable of dynamically manipulating electromagnetic waves are an emerging field in memories, optical modulators, and thermal management. Recently, their multispectral design preliminarily attracts much attention, aiming to enhance their efficiency and integration of functionalities. However, the multispectral manipulation based on these materials is challenging due to their ubiquitous wavelength dependence restricting their capacity to narrow wavelengths. In this article, we cascade multiple tunable optical cavities with selective-transparent layers, enabling a universal approach to overcoming wavelength dependence and establishing a multispectral platform with highly integrated functions. Based on it, we demonstrate the multispectral (ranging from 400 nm to 3 cm), fast response speed (0.9 s), and reversible manipulation based on a typical phase change material, vanadium dioxide. Our platform involves tandem VO2-based Fabry-Pérot (F-P) cavities enabling the customization of optical responses at target bands independently. It can achieve broadband color-changing capacity in the visible region (a shift of ~60 nm in resonant wavelength) and is capable of freely switching between three typical optical models (transmittance, reflectance, and absorptance) in the infrared to microwave regions with drastic amplitude tunability exceeding 0.7. This work represents a state-of-art advance in multispectral optics and material science, providing a critical approach for expanding the multispectral manipulation ability of optical systems.

LINC01234 Sponging of the miR-513a-5p/AOX1 Axis is Upregulated in Osteoporosis and Regulates Osteogenic Differentiation of Bone Marrow Mesenchymal Stem Cells.

Non-coding RNAs, including long-chain non-coding RNA (lncRNA) and micro-RNA (miRNA), have been implicated in osteoporosis (OP) progression by regulating osteoblast-dependent bone metabolism. Herein, we investigated whether LINC01234, miR-513a-5p, and AOX1 regulate osteogenic differentiation and proliferation of human bone marrow mesenchymal stem cells (hMSCs). The expression of LINC01234, miR-513a-5p, and AOX1 was monitored using RT-qPCR or western blotting. Cell proliferation was assessed using a CCK8 assay. Alkaline phosphatase activity (ALP) and alizarin red dye staining were performed to determine osteogenic differentiation. Furthermore, the expression of osteoblast differentiation markers, such as ALP, BMP1 (bone morphogenetic protein 1), osteocalcin (OCN), and osteopontin (OPN), was determined by RT-qPCR. Luciferase reporter and RNA immunoprecipitation (RIP) assays were performed to verify the interplay between miR-513a-5p and LINC01234 or AOX1. Compared with the plasma of healthy controls, LINC01234 and AOX1 were highly expressed in the plasma of patients with OP, whereas miR-513a-5p showed low expression. In contrast, LINC01234 and AOX1 expression displayed a gradual decrease in induced differentiated hMSCs, while miR-513a-5p expression was upregulated with induction time. The predicted binding sites between miR-513a-5p and LINC01234 or AOX1 were verified by luciferase reporter and RIP assays. LINC01234 silencing induced osteogenic differentiation and proliferation in vitro, and miR-513a-5p silencing blunted osteogenic differentiation and proliferation modulated by LINC01234. AOX1 silencing caused by miR-513a-5p enhances osteogenic proliferation and differentiation. LINC01234 sponging of the miR-513a-5p/AOX1 axis impeded the osteogenic differentiation of hMSCs, favoring OP progression.

Multi-Modal Remote Sensing Image Segmentation with Intuition-Inspired Hypergraph Modeling.

Multi-modal remote sensing (RS) image segmentation aims to comprehensively utilize multiple RS modalities to assign pixel-level semantics to the studied scenes, which can provide a new perspective for global city understanding. Multi-modal segmentation inevitably encounters the challenge of modeling intra- and inter-modal relationships, i.e., object diversity and modal gaps. However, the previous methods are usually designed for a single RS modality, limited by the noisy collection environment and poor discrimination information. Neuropsychology and neuroanatomy confirm that the human brain performs the guiding perception and integrative cognition of multi-modal semantics through intuitive reasoning. Therefore, establishing a semantic understanding framework inspired by intuition to realize multi-modal RS segmentation becomes the main motivation of this work. Drived by the superiority of hypergraphs in modeling high-order relationships, we propose an intuition-inspired hypergraph network (I2HN) for multi-modal RS segmentation. Specifically, we present a hypergraph parser to imitate guiding perception to learn intra-modal object-wise relationships. It parses the input modality into irregular hypergraphs to mine semantic clues and generate robust mono-modal representations. In addition, we also design a hypergraph matcher to dynamically update the hypergraph structure from the explicit correspondence of visual concepts, similar to integrative cognition, to improve cross-modal compatibility when fusing multi-modal features. Extensive experiments on two multi-modal RS datasets show that the proposed I2HN outperforms the state-of-the-art models, achieving F1/mIoU accuracy 91.4%/82.9% on the ISPRS Vaihingen dataset, and 92.1%/84.2% on the MSAW dataset. The complete algorithm and benchmark results will be available online.

TGF-ß score based on silico analysis can robustly predict prognosis and immunological characteristics in lower-grade glioma: the evidence from multicenter studies.

INTRODUCTION: Nowadays, mounting evidence shows that variations in TGF-ß signaling pathway-related components influence tumor development. Current research has patents describing the use of anti-TGF-ß antibodies and checkpoint inhibitors for the treatment of proliferative diseases. Importantly, TGF-ß signaling pathway is significant for lower-grade glioma (LGG) to evade host immunity. Loss of particular tumor antigens and shutdown of professional antigen-presenting cell activity may render the anti-tumor response ineffective in LGG patients. However, the prognostic significance of TGF-ß related genes in LGG is still unknown. METHODS: We collected RNA-seq data from the GTEx database (normal cortical tissues), the Cancer Genome Atlas database (TCGA-LGG), and the Chinese Glioma Genome Atlas database (CGGA-693 and CGGA-325) for conducting our investigation. RESULTS: In addition, previous publications were explored for the 223 regulators of the TGF-ß signaling pathway, and 30 regulators with abnormal expression in TCGA and GTEx database were identified. In order to identify hub prognostic regulators, least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox regression analysis were used to screen from differentially expressed genes (DEGs). On the basis of 11 genes from LASSO-Cox regression analysis (NEDD8, CHRD, TGFBR1, TP53, BMP2, LRRC32, THBS2, ID1, NOG, TNF, and SERPINE1), TGF-ß score was calculated. Multiple statistical approaches verified the predictive value of the TGF-ß score for the training cohort and two external validation cohorts. Considering the importance of the TGF-ß signaling pathway in immune regulation, we evaluated the prediction of the TGF-ß score for immunological characteristics and the possible application of the immunotherapeutic response using six algorithms (TIMER, CIBERSORT, QUANTISEQ, MCP-counter, XCELL and EPIC) and three immunotherapy cohorts (GSE78820, Imvigor-210 and PRJEB23709). Notably, we compared our risk signature with the signature in ten publications in the meta-cohort (TCGA-LGG, CGGA-693 and CGGA-325), and the TGF-ß score had the best predictive efficiency (C-index =0.812). CONCLUSION: In conclusion, our findings suggest that TGF-ß signaling pathway-related signatures are prognostic biomarkers in LGG and provide a novel tool for tumor microenvironment (TME) assessment.

A novel graphene/triolein complex-based lubricant for improving high temperature water-based drilling fluid.

Drilling engineering plays a pivotal role in the exploration and extraction of subsurface resources. It heavily depends on drilling fluid, which serves various essential functions including cooling the drill bit, removing drilled cuttings, maintaining formation pressure equilibrium, stabilizing the wellbore, transmitting hydraulic pressure, and safeguarding oil and gas reservoirs. Nonetheless, drilling fluid encounters multiple obstacles such as leakage control, waste fluid management, prevention of wellbore collapse, avoidance of hole enlargement, and environmental preservation. In order to surmount these challenges, the introduction of lubricants into the drilling fluid yields a multitude of advantages, encompassing equipment safeguarding, enhanced drilling efficiency, preservation of wellbore integrity, and bolstered drilling safety. These factors hold crucial significance in ensuring the triumph of drilling operations. This paper presents the introduction of a new lubricant derived from triolein. Following the preparation of graphene and triolein, they were incorporated into the drilling fluid system. A set of tests was subsequently conducted after aging at 240 °C for 16 hours. To assess the impact of the lubricant on the drilling fluid, an examination of rheological and filtration properties was conducted. Additionally, investigations into the friction coefficient, adhesion coefficient, and extreme pressure lubricity were carried out to evaluate the lubricating performance of the drilling fluid. Adding lubricants at a temperature of 240 degrees Celsius has successfully controlled the adhesion coefficient of the drilling fluid to below 0.2, reaching a minimum of 0.055, resulting in a reduction rate of over 70%. This indicates that the lubricant performs well at high temperatures, effectively reducing friction and enhancing drilling speed.

Comparison of lipid structure and composition in human or cow's milk with different fat globules by homogenization.

Homogenization is used in human milk to add supplements for premature infants and in cow's milk to make it more uniform and stable for commercial purposes. However, it may destroy the milk fat globule (MFG) structure and composition, affecting its functional characteristics. This study aims to compare human and cow's milk with particle size ranges of 4-6 µm (large-sized), 1-2 µm (medium-sized), and 0.3-0.5 µm (small-sized) before and after homogenization at different pressure levels. CLSM and SDS-PAGE were used to perform the structural characterization. The lipid compositions were analyzed using GC and LC-MS. The results showed that homogenization obviously changed the MFG structure and its lipid composition. After homogenization, more caseins and whey proteins were adsorbed on both the human and cow's milk fat globule interface, while the proteins observed in human milk were dispersed. This could be attributed to the different types and contents of proteins initially. The influence of homogenization on milk phospholipids was higher than triacylglycerols and fatty acids, which was highly correlated with their initial distributions in MFGs. These results provide new information about the interfacial composition of human and cow's milk fat globules upon homogenization and establish the scientific basis for homogenization application in human and cow's milk to help explore their potential functions.

Automated object recognition in high-resolution optical remote sensing imagery.

This paper reports the background and results of the Automated Object Recognition in Optical Remote Sensing Imagery, which is one of the tracks in 2022 International Algorithm Case Competition, as well as summarize the challenges, champion solutions, and future directions.

Revealing influencing factors on global waste distribution via deep-learning based dumpsite detection from satellite imagery.

With the advancement of global civilisation, monitoring and managing dumpsites have become essential parts of environmental governance in various countries. Dumpsite locations are difficult to obtain in a timely manner by local government agencies and environmental groups. The World Bank shows that governments need to spend massive labour and economic costs to collect illegal dumpsites to implement management. Here we show that applying novel deep convolutional networks to high-resolution satellite images can provide an effective, efficient, and low-cost method to detect dumpsites. In sampled areas of 28 cities around the world, our model detects nearly 1000 dumpsites that appeared around 2021. This approach reduces the investigation time by more than 96.8% compared with the manual method. With this novel and powerful methodology, it is now capable of analysing the relationship between dumpsites and various social attributes on a global scale, temporally and spatially.

Multicolor Light Mixing in Optofluidic Concave Interfaces for Anticounterfeiting with Deep Learning Authentication.

Anticounterfeiting technology has received tremendous interest for its significance in daily necessities, medical industry, and high-end products. Confidential tags based on photoluminescence are one of the most widely used approaches for their vivid visualization and high throughput. However, the complexity of confidential tags is generally limited to the accessibility of inks and their spatial location; generating an infinite combination of emission colors is therefore a challenging task. Here, we demonstrate a concept to create complex color light mixing in a confined space formed by microscale optofluidic concave interfaces. Infinite color combination and capacity were generated through chaotic behavior of light mixing and interaction in an ininkjet-printed skydome structure. Through the chaotic mixing of emission intensity, wavelength, and light propagation trajectories, the visionary patterns serve as a highly unclonable label. Finally, a deep learning-based machine vision system was built for the authentication process. The developed anticounterfeiting system may provide inspiration for utilizing space color mixing in optical security and communication applications.