CauDR: A causality-inspired domain generalization framework for fundus-based diabetic retinopathy grading.

Diabetic retinopathy (DR) is the most common diabetic complication, which usually leads to retinal damage, vision loss, and even blindness. A computer-aided DR grading system has a significant impact on helping ophthalmologists with rapid screening and diagnosis. Recent advances in fundus photography have precipitated the development of novel retinal imaging cameras and their subsequent implementation in clinical practice. However, most deep learning-based algorithms for DR grading demonstrate limited generalization across domains. This inferior performance stems from variance in imaging protocols and devices inducing domain shifts. We posit that declining model performance between domains arises from learning spurious correlations in the data. Incorporating do-operations from causality analysis into model architectures may mitigate this issue and improve generalizability. Specifically, a novel universal structural causal model (SCM) was proposed to analyze spurious correlations in fundus imaging. Building on this, a causality-inspired diabetic retinopathy grading framework named CauDR was developed to eliminate spurious correlations and achieve more generalizable DR diagnostics. Furthermore, existing datasets were reorganized into 4DR benchmark for DG scenario. Results demonstrate the effectiveness and the state-of-the-art (SOTA) performance of CauDR. Diabetic retinopathy (DR) is the most common diabetic complication, which usually leads to retinal damage, vision loss, and even blindness. A computer-aided DR grading system has a significant impact on helping ophthalmologists with rapid screening and diagnosis. Recent advances in fundus photography have precipitated the development of novel retinal imaging cameras and their subsequent implementation in clinical practice. However, most deep learning-based algorithms for DR grading demonstrate limited generalization across domains. This inferior performance stems from variance in imaging protocols and devices inducing domain shifts. We posit that declining model performance between domains arises from learning spurious correlations in the data. Incorporating do-operations from causality analysis into model architectures may mitigate this issue and improve generalizability. Specifically, a novel universal structural causal model (SCM) was proposed to analyze spurious correlations in fundus imaging. Building on this, a causality-inspired diabetic retinopathy grading framework named CauDR was developed to eliminate spurious correlations and achieve more generalizable DR diagnostics. Furthermore, existing datasets were reorganized into 4DR benchmark for DG scenario. Results demonstrate the effectiveness and the state-of-the-art (SOTA) performance of CauDR.

CoT: a transformer-based method for inferring tumor clonal copy number substructure from scDNA-seq data.

Single-cell DNA sequencing (scDNA-seq) has been an effective means to unscramble intra-tumor heterogeneity, while joint inference of tumor clones and their respective copy number profiles remains a challenging task due to the noisy nature of scDNA-seq data. We introduce a new bioinformatics method called CoT for deciphering clonal copy number substructure. The backbone of CoT is a Copy number Transformer autoencoder that leverages multi-head attention mechanism to explore correlations between different genomic regions, and thus capture global features to create latent embeddings for the cells. CoT makes it convenient to first infer cell subpopulations based on the learned embeddings, and then estimate single-cell copy numbers through joint analysis of read counts data for the cells belonging to the same cluster. This exploitation of clonal substructure information in copy number analysis helps to alleviate the effect of read counts non-uniformity, and yield robust estimations of the tumor copy numbers. Performance evaluation on synthetic and real datasets showcases that CoT outperforms the state of the arts, and is highly useful for deciphering clonal copy number substructure.

Vitamin C Inhibited Pulmonary Metastasis through Activating Nrf2/HO-1 Pathway.

As an important nutritional component, vitamin C (Vc) shows good antitumor activity in a variety of cancer, but there are few studies in pulmonary metastasis. In order to verify its anticancer and antimetastatic effect, the study sets up H22 pulmonary metastasis mouse model. The results show that intraperitoneal injection of Vc inhibits pulmonary metastasis through up-regulating the expression of Nrf2, HO-1, cleaved caspases 3 and 9, and causing DNA damage and apoptosis which is similar to the pro-oxidant effect of Vc in p53 null cells (H1299 cells). Meanwhile, oral administration of Vc up-regulates the expression of p53, directly activates Nrf2/HO-1 pathway, increases expression of cleaved caspases 3 and 9, and ultimately inhibits pulmonary metastasis, which is the same as the antioxidant result of Vc in p53 wild-type cells. In addition, Vc inhibits the proliferation and migration of lung cancer cells in a concentration-dependent manner and has little cytotoxic effects on normal cells. Notably, the experiment further illustrates that besides intravenous Vc, oral Vc significantly inhibits the pulmonary metastasis in mice. All in all, these findings provide new clues for Vc-treated pulmonary metastasis in clinical research.

Inferring single-cell copy number profiles through cross-cell segmentation of read counts.

BACKGROUND: Copy number alteration (CNA) is one of the major genomic variations that frequently occur in cancers, and accurate inference of CNAs is essential for unmasking intra-tumor heterogeneity (ITH) and tumor evolutionary history. Single-cell DNA sequencing (scDNA-seq) makes it convenient to profile CNAs at single-cell resolution, and thus aids in better characterization of ITH. Despite that several computational methods have been proposed to decipher single-cell CNAs, their performance is limited in either breakpoint detection or copy number estimation due to the high dimensionality and noisy nature of read counts data. RESULTS: By treating breakpoint detection as a process to segment high dimensional read count sequence, we develop a novel method called DeepCNA for cross-cell segmentation of read count sequence and per-cell inference of CNAs. To cope with the difficulty of segmentation, an autoencoder (AE) network is employed in DeepCNA to project the original data into a low-dimensional space, where the breakpoints can be efficiently detected along each latent dimension and further merged to obtain the final breakpoints. Unlike the existing methods that manually calculate certain statistics of read counts to find breakpoints, the AE model makes it convenient to automatically learn the representations. Based on the inferred breakpoints, we employ a mixture model to predict copy numbers of segments for each cell, and leverage expectation-maximization algorithm to efficiently estimate cell ploidy by exploring the most abundant copy number state. Benchmarking results on simulated and real data demonstrate our method is able to accurately infer breakpoints as well as absolute copy numbers and surpasses the existing methods under different test conditions. DeepCNA can be accessed at: https://github.com/zhyu-lab/deepcna . CONCLUSIONS: Profiling single-cell CNAs based on deep learning is becoming a new paradigm of scDNA-seq data analysis, and DeepCNA is an enhancement to the current arsenal of computational methods for investigating cancer genomics.

rcCAE: a convolutional autoencoder method for detecting intra-tumor heterogeneity and single-cell copy number alterations.

Intra-tumor heterogeneity (ITH) is one of the major confounding factors that result in cancer relapse, and deciphering ITH is essential for personalized therapy. Single-cell DNA sequencing (scDNA-seq) now enables profiling of single-cell copy number alterations (CNAs) and thus aids in high-resolution inference of ITH. Here, we introduce an integrated framework called rcCAE to accurately infer cell subpopulations and single-cell CNAs from scDNA-seq data. A convolutional autoencoder (CAE) is employed in rcCAE to learn latent representation of the cells as well as distill copy number information from noisy read counts data. This unsupervised representation learning via the CAE model makes it convenient to accurately cluster cells over the low-dimensional latent space, and detect single-cell CNAs from enhanced read counts data. Extensive performance evaluations on simulated datasets show that rcCAE outperforms the existing CNA calling methods, and is highly effective in inferring clonal architecture. Furthermore, evaluations of rcCAE on two real datasets demonstrate that it is able to provide a more refined clonal structure, of which some details are lost in clonal inference based on integer copy numbers.

Efficient and accurate large library ligand docking with KarmaDock.

Ligand docking is one of the core technologies in structure-based virtual screening for drug discovery. However, conventional docking tools and existing deep learning tools may suffer from limited performance in terms of speed, pose quality and binding affinity accuracy. Here we propose KarmaDock, a deep learning approach for ligand docking that integrates the functions of docking acceleration, binding pose generation and correction, and binding strength estimation. The three-stage model consists of the following components: (1) encoders for the protein and ligand to learn the representations of intramolecular interactions; (2) E(n) equivariant graph neural networks with self-attention to update the ligand pose based on both protein-ligand and intramolecular interactions, followed by post-processing to ensure chemically plausible structures; (3) a mixture density network for scoring the binding strength. KarmaDock was validated on four benchmark datasets and tested in a real-world virtual screening project that successfully identified experiment-validated active inhibitors of leukocyte tyrosine kinase (LTK).

Contrastive-ACE : Domain Generalization Through Alignment of Causal Mechanisms.

Domain generalization aims to learn knowledge invariant across different distributions while semantically meaningful for downstream tasks from multiple source domains, to improve the model's generalization ability on unseen target domains. The fundamental objective is to understand the underlying "invariance" behind these observational distributions and such invariance has been shown to have a close connection to causality. While many existing approaches make use of the property that causal features are invariant across domains, we consider the invariance of the average causal effect of the features to the labels. This invariance regularizes our training approach in which interventions are performed on features to enforce stability of the causal prediction by the classifier across domains. Our work thus sheds some light on the domain generalization problem by introducing invariance of the mechanisms into the learning process. Experiments on several benchmark datasets demonstrate the performance of the proposed method against SOTAs.

Paris saponin II-induced paraptosis-associated cell death increased the sensitivity of cisplatin.

Paris Saponin II (PSII) has been regarded as an effective and imperative component isolated from Rhizoma Paridis saponins (RPS) and exhibited strong anti-tumor effects on a variety of cancer. Our results revealed that human non-small lung cancer cell lines NCI-H460 and NCI-H520 were exposed to 1 µM of PSII, which inhibited the proliferation of lung cancer cells and activated apoptosis, autophagy and paraptosis. PSII induced paraptosis-associated cell death prior to apoptosis and autophagy. It induced paraptosis based on ER stress through activation of the JNK pathway. Meanwhile, PSII increased the cytotoxicity of cisplatin through paraptosis-associated pathway. All in all, PSII induced paraptosis based on induction of non-apoptotic cell death, which would be a possible approach to suppress the multi-drug resistant to apoptosis.

Confounder Detection in High-Dimensional Linear Models Using First Moments of Spectral Measures.

In this letter, we study the confounder detection problem in the linear model, where the target variable [Formula: see text] is predicted using its [Formula: see text] potential causes [Formula: see text]. Based on an assumption of a rotation-invariant generating process of the model, recent study shows that the spectral measure induced by the regression coefficient vector with respect to the covariance matrix of [Formula: see text] is close to a uniform measure in purely causal cases, but it differs from a uniform measure characteristically in the presence of a scalar confounder. Analyzing spectral measure patterns could help to detect confounding. In this letter, we propose to use the first moment of the spectral measure for confounder detection. We calculate the first moment of the regression vector-induced spectral measure and compare it with the first moment of a uniform spectral measure, both defined with respect to the covariance matrix of [Formula: see text]. The two moments coincide in nonconfounding cases and differ from each other in the presence of confounding. This statistical causal-confounding asymmetry can be used for confounder detection. Without the need to analyze the spectral measure pattern, our method avoids the difficulty of metric choice and multiple parameter optimization. Experiments on synthetic and real data show the performance of this method.

Cellulose-dependent expression and antibacterial characteristics of surfactin from Bacillus subtilis HH2 isolated from the giant panda.

Surfactin secreted by Bacillus subtilis can confer strong, diverse antipathogenic effects, thereby benefitting the host. Carbon source is an important factor for surfactin production. However, the mechanism that bacteria utilize cellulose, the most abundant substance in the intestines of herbivores, to produce surfactin remains unclear. Here, we used B. subtilis HH2, isolated from the feces of a giant panda, as a model to determine changes in surfactin expression in the presence of different concentrations of cellulose by quantitative polymerase chain reaction and high-performance liquid chromatography. We further investigated the antimicrobial effects of surfactin against three common intestinal pathogens (Escherichia coli, Staphylococcus aureus, and Salmonella enterica) and its resistance to high temperature (60-121°C), pH (1-12), trypsin (100-300 µg/mL, pH 8), and pepsin (100-300 µg/mL, pH 2). The results showed that the surfactin expressed lowest in bacteria cultured in the presence of 1% glucose medium as the carbon source, whereas increased in an appropriate cellulose concentration (0.67% glucose and 0.33% cellulose). The surfactin could inhibit E. coli and Staphylococcus aureus, but did not affect efficiently for Salmonella enterica. The antibacterial ability of surfactin did not differ according to temperature (60-100°C), pH (2-11), trypsin (100-300 µg/mL), and pepsin (100-300 µg/mL; P > 0.05), but decreased significantly at extreme environments (121°C, pH 1 or 12; P < 0.05) compared with that in the control group (37°C, pH = 7, without any protease). In conclusion, our findings indicated that B. subtilis HH2 could increase surfactin expression in an appropriate cellulose environment and thus provide benefits to improve the intestinal health of herbivores.

Causal Inference on Multidimensional Data Using Free Probability Theory.

In this paper, we deal with the problem of inferring causal relations for multidimensional data. Based on the postulate that the distribution of the cause and the conditional distribution of the effect given cause are generated independently, we show that the covariance matrix of the mean embedding of the cause in reproducing kernel Hilbert space (RKHS) is free independent with the covariance matrix of the conditional embedding of the effect given cause. This, called freeness condition, induces a cause-effect asymmetry that a designed measurement is 0 in the causal direction but smaller than 0 in the anticausal direction, and it uncovers the causal direction. One important novel aspect of this paper is that we interpret the independence as a freeness condition between covariance matrices of RKHS distribution embeddings, and it has a wide applicability. We show that our freeness condition-based inference method succeeds in scenarios like additive noise cases, where other methods fail, by theoretical analysis and experimental results.

Causal Inference on Discrete Data via Estimating Distance Correlations.

In this article, we deal with the problem of inferring causal directions when the data are on discrete domain. By considering the distribution of the cause [Formula: see text] and the conditional distribution mapping cause to effect [Formula: see text] as independent random variables, we propose to infer the causal direction by comparing the distance correlation between [Formula: see text] and [Formula: see text] with the distance correlation between [Formula: see text] and [Formula: see text]. We infer that X causes Y if the dependence coefficient between [Formula: see text] and [Formula: see text] is smaller. Experiments are performed to show the performance of the proposed method.

Transcriptional regulation and adaptation to a high-fiber environment in Bacillus subtilis HH2 isolated from feces of the giant panda.

In the giant panda, adaptation to a high-fiber environment is a first step for the adequate functioning of intestinal bacteria, as the high cellulose content of the gut due to the panda's vegetarian appetite results in a harsh environment. As an excellent producer of several enzymes and vitamins, Bacillus subtilis imparts various advantages to animals. In our previous study, we determined that several strains of B. subtilis isolated from pandas exhibited good cellulose decomposition ability, and we hypothesized that this bacterial species can survive in and adapt well to a high-fiber environment. To evaluate this hypothesis, we employed RNA-Seq technology to analyze the differentially expressed genes of the selected strain B. subtilis HH2, which demonstrates significant cellulose hydrolysis of different carbon sources (cellulose and glucose). In addition, we used bioinformatics software and resources to analyze the functions and pathways of differentially expressed genes. Interestingly, comparison of the cellulose and glucose groups revealed that the up-regulated genes were involved in amino acid and lipid metabolism or transmembrane transport, both of which are involved in cellulose utilization. Conversely, the down-regulated genes were involved in non-essential functions for bacterial life, such as toxin and bacteriocin secretion, possibly to conserve energy for environmental adaptation. The results indicate that B. subtilis HH2 triggered a series of adaptive mechanisms at the transcriptional level, which suggests that this bacterium could act as a probiotic for pandas fed a high-fiber diet, despite the fact that cellulose is not a very suitable carbon source for this bacterial species. In this study, we present a model to understand the dynamic organization of and interactions between various functional and regulatory networks for unicellular organisms in a high-fiber environment.