CNS-Native Myeloid Cells Drive Immune Suppression in the Brain Metastatic Niche through Cxcl10.

Brain metastasis (br-met) develops in an immunologically unique br-met niche. Central nervous system-native myeloid cells (CNS-myeloids) and bone-marrow-derived myeloid cells (BMDMs) cooperatively regulate brain immunity. The phenotypic heterogeneity and specific roles of these myeloid subsets in shaping the br-met niche to regulate br-met outgrowth have not been fully revealed. Applying multimodal single-cell analyses, we elucidated a heterogeneous but spatially defined CNS-myeloid response during br-met outgrowth. We found Ccr2+ BMDMs minimally influenced br-met while CNS-myeloid promoted br-met outgrowth. Additionally, br-met-associated CNS-myeloid exhibited downregulation of Cx3cr1. Cx3cr1 knockout in CNS-myeloid increased br-met incidence, leading to an enriched interferon response signature and Cxcl10 upregulation. Significantly, neutralization of Cxcl10 reduced br-met, while rCxcl10 increased br-met and recruited VISTAHi PD-L1+ CNS-myeloid to br-met lesions. Inhibiting VISTA- and PD-L1-signaling relieved immune suppression and reduced br-met burden. Our results demonstrate that loss of Cx3cr1 in CNS-myeloid triggers a Cxcl10-mediated vicious cycle, cultivating a br-met-promoting, immune-suppressive niche.

TGSA: protein-protein association-based twin graph neural networks for drug response prediction with similarity augmentation.

MOTIVATION: Drug response prediction (DRP) plays an important role in precision medicine (e.g. for cancer analysis and treatment). Recent advances in deep learning algorithms make it possible to predict drug responses accurately based on genetic profiles. However, existing methods ignore the potential relationships among genes. In addition, similarity among cell lines/drugs was rarely considered explicitly. RESULTS: We propose a novel DRP framework, called TGSA, to make better use of prior domain knowledge. TGSA consists of Twin Graph neural networks for Drug Response Prediction (TGDRP) and a Similarity Augmentation (SA) module to fuse fine-grained and coarse-grained information. Specifically, TGDRP abstracts cell lines as graphs based on STRING protein-protein association networks and uses Graph Neural Networks (GNNs) for representation learning. SA views DRP as an edge regression problem on a heterogeneous graph and utilizes GNNs to smooth the representations of similar cell lines/drugs. Besides, we introduce an auxiliary pre-training strategy to remedy the identified limitations of scarce data and poor out-of-distribution generalization. Extensive experiments on the GDSC2 dataset demonstrate that our TGSA consistently outperforms all the state-of-the-art baselines under various experimental settings. We further evaluate the effectiveness and contributions of each component of TGSA via ablation experiments. The promising performance of TGSA shows enormous potential for clinical applications in precision medicine. AVAILABILITY AND IMPLEMENTATION: The source code is available at https://github.com/violet-sto/TGSA. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Globally optimal OCT surface segmentation using a constrained IPM optimization.

Segmentation of multiple surfaces in optical coherence tomography (OCT) images is a challenging problem, further complicated by the frequent presence of weak boundaries, varying layer thicknesses, and mutual influence between adjacent surfaces. The traditional graph-based optimal surface segmentation method has proven its effectiveness with its ability to capture various surface priors in a uniform graph model. However, its efficacy heavily relies on handcrafted features that are used to define the surface cost for the "goodness" of a surface. Recently, deep learning (DL) is emerging as a powerful tool for medical image segmentation thanks to its superior feature learning capability. Unfortunately, due to the scarcity of training data in medical imaging, it is nontrivial for DL networks to implicitly learn the global structure of the target surfaces, including surface interactions. This study proposes to parameterize the surface cost functions in the graph model and leverage DL to learn those parameters. The multiple optimal surfaces are then simultaneously detected by minimizing the total surface cost while explicitly enforcing the mutual surface interaction constraints. The optimization problem is solved by the primal-dual interior-point method (IPM), which can be implemented by a layer of neural networks, enabling efficient end-to-end training of the whole network. Experiments on spectral-domain optical coherence tomography (SD-OCT) retinal layer segmentation demonstrated promising segmentation results with sub-pixel accuracy.

Label-free visualization and characterization of extracellular vesicles in breast cancer.

Despite extensive interest, extracellular vesicle (EV) research remains technically challenging. One of the unexplored gaps in EV research has been the inability to characterize the spatially and functionally heterogeneous populations of EVs based on their metabolic profile. In this paper, we utilize the intrinsic optical metabolic and structural contrast of EVs and demonstrate in vivo/in situ characterization of EVs in a variety of unprocessed (pre)clinical samples. With a pixel-level segmentation mask provided by the deep neural network, individual EVs can be analyzed in terms of their optical signature in the context of their spatial distribution. Quantitative analysis of living tumor-bearing animals and fresh excised human breast tissue revealed abundance of NAD(P)H-rich EVs within the tumor, near the tumor boundary, and around vessel structures. Furthermore, the percentage of NAD(P)H-rich EVs is highly correlated with human breast cancer diagnosis, which emphasizes the important role of metabolic imaging for EV characterization as well as its potential for clinical applications. In addition to the characterization of EV properties, we also demonstrate label-free monitoring of EV dynamics (uptake, release, and movement) in live cells and animals. The in situ metabolic profiling capacity of the proposed method together with the finding of increasing NAD(P)H-rich EV subpopulations in breast cancer have the potential for empowering applications in basic science and enhancing our understanding of the active metabolic roles that EVs play in cancer progression.

Three-dimensional visualization and a deep-learning model reveal complex fungal parasite networks in behaviorally manipulated ants.

Some microbes possess the ability to adaptively manipulate host behavior. To better understand how such microbial parasites control animal behavior, we examine the cell-level interactions between the species-specific fungal parasite Ophiocordyceps unilateralis sensu lato and its carpenter ant host (Camponotus castaneus) at a crucial moment in the parasite's lifecycle: when the manipulated host fixes itself permanently to a substrate by its mandibles. The fungus is known to secrete tissue-specific metabolites and cause changes in host gene expression as well as atrophy in the mandible muscles of its ant host, but it is unknown how the fungus coordinates these effects to manipulate its host's behavior. In this study, we combine techniques in serial block-face scanning-electron microscopy and deep-learning-based image segmentation algorithms to visualize the distribution, abundance, and interactions of this fungus inside the body of its manipulated host. Fungal cells were found throughout the host body but not in the brain, implying that behavioral control of the animal body by this microbe occurs peripherally. Additionally, fungal cells invaded host muscle fibers and joined together to form networks that encircled the muscles. These networks may represent a collective foraging behavior of this parasite, which may in turn facilitate host manipulation.

Decoding Calcium Signaling Dynamics during Drosophila Wing Disc Development.

The robust specification of organ development depends on coordinated cell-cell communication. This process requires signal integration among multiple pathways, relying on second messengers such as calcium ions. Calcium signaling encodes a significant portion of the cellular state by regulating transcription factors, enzymes, and cytoskeletal proteins. However, the relationships between the inputs specifying cell and organ development, calcium signaling dynamics, and final organ morphology are poorly understood. Here, we have designed a quantitative image-analysis pipeline for decoding organ-level calcium signaling. With this pipeline, we extracted spatiotemporal features of calcium signaling dynamics during the development of the Drosophila larval wing disc, a genetic model for organogenesis. We identified specific classes of wing phenotypes that resulted from calcium signaling pathway perturbations, including defects in gross morphology, vein differentiation, and overall size. We found four qualitative classes of calcium signaling activity. These classes can be ordered based on agonist stimulation strength Gαq-mediated signaling. In vivo calcium signaling dynamics depend on both receptor tyrosine kinase/phospholipase C γ and G protein-coupled receptor/phospholipase C ß activities. We found that spatially patterned calcium dynamics correlate with known differential growth rates between anterior and posterior compartments. Integrated calcium signaling activity decreases with increasing tissue size, and it responds to morphogenetic perturbations that impact organ growth. Together, these findings define how calcium signaling dynamics integrate upstream inputs to mediate multiple response outputs in developing epithelial organs.

Single Cells Exhibit Differing Behavioral Phases during Early Stages of Pseudomonas aeruginosa Swarming.

Pseudomonas aeruginosa is among the many bacteria that swarm, where groups of cells coordinate to move over surfaces. It has been challenging to determine the behavior of single cells within these high-cell-density swarms. To track individual cells within P. aeruginosa swarms, we imaged a fluorescently labeled subset of the larger population. Single cells at the advancing swarm edge varied in their motility dynamics as a function of time. From these data, we delineated four phases of early swarming prior to the formation of the tendril fractals characteristic of P. aeruginosa swarming by collectively considering both micro- and macroscale data. We determined that the period of greatest single-cell motility does not coincide with the period of greatest collective swarm expansion. We also noted that flagellar, rhamnolipid, and type IV pilus motility mutants exhibit substantially less single-cell motility than the wild type.IMPORTANCE Numerous bacteria exhibit coordinated swarming motion over surfaces. It is often challenging to assess the behavior of single cells within swarming communities due to the limitations of identifying, tracking, and analyzing the traits of swarming cells over time. Here, we show that the behavior of Pseudomonas aeruginosa swarming cells can vary substantially in the earliest phases of swarming. This is important to establish that dynamic behaviors should not be assumed to be constant over long periods when predicting and simulating the actions of swarming bacteria.

Single molecule sequencing-guided scaffolding and correction of draft assemblies.

BACKGROUND: Although single molecule sequencing is still improving, the lengths of the generated sequences are inevitably an advantage in genome assembly. Prior work that utilizes long reads to conduct genome assembly has mostly focused on correcting sequencing errors and improving contiguity of de novo assemblies. RESULTS: We propose a disassembling-reassembling approach for both correcting structural errors in the draft assembly and scaffolding a target assembly based on error-corrected single molecule sequences. To achieve this goal, we formulate a maximum alternating path cover problem. We prove that this problem is NP-hard, and solve it by a 2-approximation algorithm. CONCLUSIONS: Our experimental results show that our approach can improve the structural correctness of target assemblies in the cost of some contiguity, even with smaller amounts of long reads. In addition, our reassembling process can also serve as a competitive scaffolder relative to well-established assembly benchmarks.

Kinetic Transition Networks for the Thomson Problem and Smale's Seventh Problem.

The Thomson problem, arrangement of identical charges on the surface of a sphere, has found many applications in physics, chemistry and biology. Here, we show that the energy landscape of the Thomson problem for N particles with N=132, 135, 138, 141, 144, 147, and 150 is single funneled, characteristic of a structure-seeking organization where the global minimum is easily accessible. Algorithmically, constructing starting points close to the global minimum of such a potential with spherical constraints is one of Smale's 18 unsolved problems in mathematics for the 21st century because it is important in the solution of univariate and bivariate random polynomial equations. By analyzing the kinetic transition networks, we show that a randomly chosen minimum is, in fact, always "close" to the global minimum in terms of the number of transition states that separate them, a characteristic of small world networks.

A seeding-searching-ensemble method for gland segmentation in H&E-stained images.

BACKGROUND: Glands are vital structures found throughout the human body and their structure and function are affected by many diseases. The ability to segment and detect glands among other types of tissues is important for the study of normal and disease processes and helps their analysis and visualization by pathologists in microscopic detail. METHODS: In this paper, we develop a new approach for segmenting and detecting intestinal glands in H&E-stained histology images, which utilizes a set of advanced image processing techniques: graph search, ensemble, feature extraction, and classification. Our method is computationally fast, preserves gland boundaries robustly and detects glands accurately. RESULTS: We tested the performance of our gland detection and segmentation method by analyzing a dataset of over 1700 glands in digitized high resolution clinical histology images obtained from normal and diseased human intestines. The experimental results show that our method outperforms considerably the state-of-the-art methods for gland segmentation and detection. CONCLUSIONS: Our method can produce high-quality segmentation and detection of non-overlapped glands that obey the natural property of glands in histology tissue images. With accurately detected and segmented glands, quantitative measurement and analysis can be developed for further studies of glands and computer-aided diagnosis.

MR2CPPIS: Accurate prediction of protein-protein interaction sites based on multi-scale Res2Net with coordinate attention mechanism.

Proteins play a vital role in various biological processes and achieve their functions through protein-protein interactions (PPIs). Thus, accurate identification of PPI sites is essential. Traditional biological methods for identifying PPIs are costly, labor-intensive, and time-consuming. The development of computational prediction methods for PPI sites offers promising alternatives. Most known deep learning (DL) methods employ layer-wise multi-scale CNNs to extract features from protein sequences. But, these methods usually neglect the spatial positions and hierarchical information embedded within protein sequences, which are actually crucial for PPI site prediction. In this paper, we propose MR2CPPIS, a novel sequence-based DL model that utilizes the multi-scale Res2Net with coordinate attention mechanism to exploit multi-scale features and enhance PPI site prediction capability. We leverage the multi-scale Res2Net to expand the receptive field for each network layer, thus capturing multi-scale information of protein sequences at a granular level. To further explore the local contextual features of each target residue, we employ a coordinate attention block to characterize the precise spatial position information, enabling the network to effectively extract long-range dependencies. We evaluate our MR2CPPIS on three public benchmark datasets (Dset 72, Dset 186, and PDBset 164), achieving state-of-the-art performance. The source codes are available at https://github.com/YyinGong/MR2CPPIS.

A Transformer-Based Knowledge Distillation Network for Cortical Cataract Grading.

Cortical cataract, a common type of cataract, is particularly difficult to be diagnosed automatically due to the complex features of the lesions. Recently, many methods based on edge detection or deep learning were proposed for automatic cataract grading. However, these methods suffer a large performance drop in cortical cataract grading due to the more complex cortical opacities and uncertain data. In this paper, we propose a novel Transformer-based Knowledge Distillation Network, called TKD-Net, for cortical cataract grading. To tackle the complex opacity problem, we first devise a zone decomposition strategy to extract more refined features and introduce special sub-scores to consider critical factors of clinical cortical opacity assessment (location, area, density) for comprehensive quantification. Next, we develop a multi-modal mix-attention Transformer to efficiently fuse sub-scores and image modality for complex feature learning. However, obtaining the sub-score modality is a challenge in the clinic, which could cause the modality missing problem instead. To simultaneously alleviate the issues of modality missing and uncertain data, we further design a Transformer-based knowledge distillation method, which uses a teacher model with perfect data to guide a student model with modality-missing and uncertain data. We conduct extensive experiments on a dataset of commonly-used slit-lamp images annotated by the LOCS III grading system to demonstrate that our TKD-Net outperforms state-of-the-art methods, as well as the effectiveness of its key components. Codes are available at https://github.com/wjh892521292/Cataract_TKD-Net.

TestFit: A plug-and-play one-pass test time method for medical image segmentation.

Deep learning (DL) based methods have been extensively studied for medical image segmentation, mostly emphasizing the design and training of DL networks. Only few attempts were made on developing methods for applying DL models in test time. In this paper, we study whether a given off-the-shelf segmentation network can be stably improved on-the-fly during test time in an online processing-and-learning fashion. We propose a new online test-time method, called TestFit, to improve results of a given off-the-shelf DL segmentation model in test time by actively fitting the test data distribution. TestFit first creates a supplementary network (SuppNet) from the given trained off-the-shelf segmentation network (this original network is referred to as OGNet) and applies SuppNet together with OGNet for test time inference. OGNet keeps its hypothesis derived from the original training set to prevent the model from collapsing, while SuppNet seeks to fit the test data distribution. Segmentation results and supervision signals (for updating SuppNet) are generated by combining the outputs of OGNet and SuppNet on the fly. TestFit needs only one pass on each test sample - the same as the traditional test model pipeline - and requires no training time preparation. Since it is challenging to look at only one test sample and no manual annotation for model update each time, we develop a series of technical treatments for improving the stability and effectiveness of our proposed online test-time training method. TestFit works in a plug-and-play fashion, requires minimal hyper-parameter tuning, and is easy to use in practice. Experiments on a large collection of 2D and 3D datasets demonstrate the capability of our TestFit method.

Congenital heart disease detection by pediatric electrocardiogram based deep learning integrated with human concepts.

Early detection is critical to achieving improved treatment outcomes for child patients with congenital heart diseases (CHDs). Therefore, developing effective CHD detection techniques using low-cost and non-invasive pediatric electrocardiogram are highly desirable. We propose a deep learning approach for CHD detection, CHDdECG, which automatically extracts features from pediatric electrocardiogram and wavelet transformation characteristics, and integrates them with key human-concept features. Developed on 65,869 cases, CHDdECG achieved ROC-AUC of 0.915 and specificity of 0.881 on a real-world test set covering 12,000 cases. Additionally, on two external test sets with 7137 and 8121 cases, the overall ROC-AUC were 0.917 and 0.907 while specificities were 0.937 and 0.907. Notably, CHDdECG surpassed cardiologists in CHD detection performance comparison, and feature importance scores suggested greater influence of automatically extracted electrocardiogram features on CHD detection compared with human-concept features, implying that CHDdECG may grasp some knowledge beyond human cognition. Our study directly impacts CHD detection with pediatric electrocardiogram and demonstrates the potential of pediatric electrocardiogram for broader benefits.

Mapping Cellular Interactions from Spatially Resolved Transcriptomics Data.

Cell-cell communication (CCC) is essential to how life forms and functions. However, accurate, high-throughput mapping of how expression of all genes in one cell affects expression of all genes in another cell is made possible only recently, through the introduction of spatially resolved transcriptomics technologies (SRTs), especially those that achieve single cell resolution. However, significant challenges remain to analyze such highly complex data properly. Here, we introduce a Bayesian multi-instance learning framework, spacia, to detect CCCs from data generated by SRTs, by uniquely exploiting their spatial modality. We highlight spacia's power to overcome fundamental limitations of popular analytical tools for inference of CCCs, including losing single-cell resolution, limited to ligand-receptor relationships and prior interaction databases, high false positive rates, and most importantly the lack of consideration of the multiple-sender-to-one-receiver paradigm. We evaluated the fitness of spacia for all three commercialized single cell resolution ST technologies: MERSCOPE/Vizgen, CosMx/Nanostring, and Xenium/10X. Spacia unveiled how endothelial cells, fibroblasts and B cells in the tumor microenvironment contribute to Epithelial-Mesenchymal Transition and lineage plasticity in prostate cancer cells. We deployed spacia in a set of pan-cancer datasets and showed that B cells also participate in PDL1/PD1 signaling in tumors. We demonstrated that a CD8+ T cell/PDL1 effectiveness signature derived from spacia analyses is associated with patient survival and response to immune checkpoint inhibitor treatments in 3,354 patients. We revealed differential spatial interaction patterns between γδ T cells and liver hepatocytes in healthy and cancerous contexts. Overall, spacia represents a notable step in advancing quantitative theories of cellular communications.

Spatial organization of dendritic cells within tumor draining lymph nodes impacts clinical outcome in breast cancer patients.

BACKGROUND: Dendritic cells (DCs) are important mediators of anti-tumor immune responses. We hypothesized that an in-depth analysis of dendritic cells and their spatial relationships to each other as well as to other immune cells within tumor draining lymph nodes (TDLNs) could provide a better understanding of immune function and dysregulation in cancer. METHODS: We analyzed immune cells within TDLNs from 59 breast cancer patients with at least 5 years of clinical follow-up using immunohistochemical staining with a novel quantitative image analysis system. We developed algorithms to analyze spatial distribution patterns of immune cells in cancer versus healthy intra-mammary lymph nodes (HLNs) to derive information about possible mechanisms underlying immune-dysregulation in breast cancer. We used the non-parametric Mann-Whitney test for inter-group comparisons, Wilcoxon Matched-Pairs Signed Ranks test for intra-group comparisons and log-rank (Mantel-Cox) test for Kaplan Maier analyses. RESULTS: Degree of clustering of DCs (in terms of spatial proximity of the cells to each other) was reduced in TDLNs compared to HLNs. While there were more numerous DC clusters in TDLNs compared to HLNs,DC clusters within TDLNs tended to have fewer member DCs and also consisted of fewer cells displaying the DC maturity marker CD83. The average number of T cells within a standardized radius of a clustered DC was increased compared to that of an unclustered DC, suggesting that DC clustering was associated with T cell interaction. Furthermore, the number of T cells within the radius of a clustered DC was reduced in tumor-positive TDLNs compared to HLNs. Importantly, clinical outcome analysis revealed that DC clustering in tumor-positive TDLNs correlated with the duration of disease-free survival in breast cancer patients. CONCLUSIONS: These findings are the first to describe the spatial organization of DCs within TDLNs and their association with survival outcome. In addition, we characterized specific changes in number, size, maturity, and T cell co-localization of such clusters. Strategies to enhance DC function in-vivo, including maturation and clustering, may provide additional tools for developing more efficacious DC cancer vaccines.

TANGO: A GO-Term Embedding Based Method for Protein Semantic Similarity Prediction.

We aim to quantitatively predict protein semantic similarities (PSS), which is vital to making biological discoveries. Previously, researchers commonly exploited Gene Ontology (GO) graphs (containing standardized hierarchically-organized GO terms for annotating distinct protein attributes) to learn GO term embeddings (vector representations) for quantifying protein attribute similarities and aggregate these embeddings to form protein embeddings for similarity measurement. However, two key properties of GO terms and annotated proteins are not yet well-explored by these learning-based methods: (1) taxonomy relations between GO terms; (2) GO terms' different contributions in describing protein semantics. In this paper, we propose TANGO, a new framework composed of a TAxoNomy-aware embedding module and an aggreGatiOn module. Our Embedding Module encodes taxonomic information into GO term embeddings by incorporating GO term topological distances in the GO graph hierarchy. Hence, distances between GO term embeddings can be used to more accurately measure shared meanings between correlated protein attributes. Our Aggregation Module automatically determines the contributions of GO terms when merging into the target protein embeddings, by mining GO term concept dependency relations in the GO graph and correlations in protein annotations. We conduct extensive experiments on several public datasets. On two PSS metrics, our new method significantly outperforms known methods by a large margin.

Identifying Electrocardiogram Abnormalities Using a Handcrafted-Rule-Enhanced Neural Network.

A large number of people suffer from life-threatening cardiac abnormalities, and electrocardiogram (ECG) analysis is beneficial to determining whether an individual is at risk of such abnormalities. Automatic ECG classification methods, especially the deep learning based ones, have been proposed to detect cardiac abnormalities using ECG records, showing good potential to improve clinical diagnosis and help early prevention of cardiovascular diseases. However, the predictions of the known neural networks still do not satisfactorily meet the needs of clinicians, and this phenomenon suggests that some information used in clinical diagnosis may not be well captured and utilized by these methods. In this paper, we introduce some rules into convolutional neural networks, which help present clinical knowledge to deep learning based ECG analysis, in order to improve automated ECG diagnosis performance. Specifically, we propose a Handcrafted-Rule-enhanced Neural Network (called HRNN) for ECG classification with standard 12-lead ECG input, which consists of a rule inference module and a deep learning module. Experiments on two large-scale public ECG datasets show that our new approach considerably outperforms existing state-of-the-art methods. Further, our proposed approach not only can improve the diagnosis performance, but also can assist in detecting mislabelled ECG samples.

Learning from algorithm-generated pseudo-annotations for detecting ants in videos.

Deep learning (DL) based detection models are powerful tools for large-scale analysis of dynamic biological behaviors in video data. Supervised training of a DL detection model often requires a large amount of manually-labeled training data which are time-consuming and labor-intensive to acquire. In this paper, we propose LFAGPA (Learn From Algorithm-Generated Pseudo-Annotations) that utilizes (noisy) annotations which are automatically generated by algorithms to train DL models for ant detection in videos. Our method consists of two main steps: (1) generate foreground objects using a (set of) state-of-the-art foreground extraction algorithm(s); (2) treat the results from step (1) as pseudo-annotations and use them to train deep neural networks for ant detection. We tackle several challenges on how to make use of automatically generated noisy annotations, how to learn from multiple annotation resources, and how to combine algorithm-generated annotations with human-labeled annotations (when available) for this learning framework. In experiments, we evaluate our method using 82 videos (totally 20,348 image frames) captured under natural conditions in a tropical rain-forest for dynamic ant behavior study. Without any manual annotation cost but only algorithm-generated annotations, our method can achieve a decent detection performance (77% in [Formula: see text] score). Moreover, when using only 10% manual annotations, our method can train a DL model to perform as well as using the full human annotations (81% in [Formula: see text] score).

A CNN-LSTM Ensemble Model for Predicting Protein-Protein Interaction Binding Sites.

Proteins commonly perform biological functions through protein-protein interactions (PPIs). The knowledge of PPI sites is imperative for the understanding of protein functions, disease mechanisms, and drug design. Traditional biological experimental methods for studying PPI sites still incur considerable drawbacks, including long experimental time and high labor costs. Therefore, many computational methods have been proposed for predicting PPI sites. However, achieving high prediction performance and overcoming severe data imbalance remain challenging issues. In this paper, we propose a new sequence-based deep learning model called CLPPIS (standing for CNN-LSTM ensemble based PPI Sites prediction). CLPPIS consists of CNN and LSTM components, which can capture spatial features and sequential features simultaneously. Further, it utilizes a novel feature group as input, which has 7 physicochemical, biophysical, and statistical properties. Besides, it adopts a batch-weighted loss function to reduce the interference of imbalance data. Our work suggests that the integration of protein spatial features and sequential features provides important information for PPI sites prediction. Evaluation on three public benchmark datasets shows that our CLPPIS model significantly outperforms existing state-of-the-art methods.