Dual Passivation Strategy for Highly Stable Blue-Luminescent CsPbBr3 Nanoplatelets.

Blue-emitting colloidal CsPbX3 (X = Br, Cl, or I) perovskite nanocrystals have emerged as one of the most fascinating materials for optoelectronic applications. However, their applicability is hindered by poor stability and a low photoluminescence efficiency. Herein, highly stable CsPbBr3 nanoplatelets exhibiting intense blue luminescence are fabricated by employing a strategy in which the morphology is regulated and the surface is subjected to dual passivation through the incorporation of zirconium acetylacetonate [Zr(acac)4]. The passivated CsPbBr3 nanocrystals exhibit adjustable light emission from green to dark blue and a controllable morphology from nanocubes (NCs) to nanoplatelets (NPLs) and nanorods accomplished by varying the content of Zr(acac)4. The optimized NPLs are characterized by a bright blue emission with a central wavelength of 459 nm and a high photoluminescence quantum yield of 90%. The addition of Zr(acac)4 in the synthesis of CsPbBr3 induces oriented growth with a two-dimensional morphology. The Zr(acac)4 can repair the surface defects of the nanocrystal surface, and the surface is also capped with the Zr(OH)4 cluster layer. Therefore, the passivated blue-emitting NPLs exhibit outstanding stability compared to that of pristine NPLs during long-term storage and exposure to light. This work provides a novel strategy for fabricating highly stable PNCs with deep-blue emission and widens their potential applications in blue-emitting optoelectronic devices.

Biomimetic nanodrug blocks CD73 to inhibit adenosine and boosts antitumor immune response synergically with photothermal stimulation.

Combination of tumor immunotherapy with photothermal therapy (PTT) is a feasible tactic to overcome the drawback of immunotherapy such as poor immune response. Via triggering the immunogenic cells death (ICD), PTT can stimulate the activity of immune cells, but meanwhile, the level of adenosine is elevated via the CD73-induced decomposition of ATP which is overexpressed accompanying with the PTT process, resulting in negative feedback to impair the immune stimulation. Herein, we developed a novel biomimetic photothermal nanodrug to specifically block CD73 for inhibition of adenosine production and more efficient priming of the suppressive immune microenvironments. The nanodrug, named as AptEM@CBA, is constructed by encapsulation of photothermal agent black phosphorus quantum dots (BPQDs) and selective CD73 inhibitor α, ß-Methyleneadenosine 5'-diphosphate (AMPCP) in chitosan nanogels, which are further covered with aptamer AS1411 modified erythrocyte membrane (EM) for biomimetic camouflage. With AS1411 induced active targeting and EM induced long blood circulation time, the enrichment of the nanodrug tumor sites is promoted. The photothermal treatment promotes the maturation of dendritic cells. Meanwhile, the release of AMPCP suppress the adenosine generation via CD73 blockade, alleviating the impairment of adenosine to dendritic cells and suppressing regulatory T cells, synergically stimulate the activity of T cells. The combination of CD73 blockade with PTT, not only suppresses the growth of primary implanted tumors, but also boosts strong antitumor immunity to inhibit the growth of distal tumors, providing good potential for tumor photoimmunotherapy.

Shell thickness-induced thermal dependence: highly sensitive core-shell CdSe/ZnS/POSS-based temperature probes.

Fluorescence nanothermometry based on quantum dots is a current research hotspot for novel non-contact temperature monitoring, and is of vital significance for the modulation and design of the sensing properties of sensors. Herein, a design strategy to modulate the temperature-sensing characteristics of quantum dots based on the thickness of a shell is proposed. In this study, CdSe/ZnS quantum dot/POSS-based temperature probe films with varying fluorescence characteristics were developed, and the influence of the ZnS shell on temperature sensing was examined by varying the thickness of the ZnS shell. The temperature dependency, linearity, range of applications, and reversibility of quantum dot thin film probes were all considerably regulated by the ZnS shell, according to research on quantum dot/POSS-based films coated with various shell thicknesses. The CdSe/ZnS temperature probe with 4 monolayers (MLs) stood out among the rest due to its strong thermal stability (at least 5 cycles), large usable temperature range (20-80 °C), and excellent temperature sensitivity (R2 > 0.994). The results demonstrated that the temperature sensing performance of quantum dots was the consequence of the combined effect of multiple temperature response properties induced by the thickness of the shell, and the shell control of quantum dots to optimize the temperature sensing performance was an essential approach for the design of temperature probes. This work demonstrates the great potential of the shell in tuning the temperature sensing performance of quantum dots and provides a viable approach for the design of quantum dot temperature probes.

Controllable fabrication of well-shaped PMBA@CsPbBr3 nanoparticles for highly sensitive detection of HCl and HBr.

We report a facile two-step strategy to construct well-shaped PMBA@CsPbBr3 nanoparticles, with this strategy involving combining in situ adsorption and controlled polymerization. The morphological evolution process and mechanism of formation of the nanoparticles were demonstrated, and the nanoparticles showed high sensitivity to corrosive acid gas. This work has provided an effective approach for fabricating well-structured perovskite-based nanocomposites.

LncFASA promotes cancer ferroptosis via modulating PRDX1 phase separation.

Ferroptosis, a unique type of non-apoptotic cell death resulting from iron-dependent lipid peroxidation, has a potential physiological function in tumor suppression, but its underlying mechanisms have not been fully elucidated. Here, we report that the long non-coding RNA (lncRNA) LncFASA increases the susceptibility of triple-negative breast cancer (TNBC) to ferroptosis. As a tumor suppressor, LncFASA drives the formation of droplets containing peroxiredoxin1 (PRDX1), a member of the peroxidase family, resulting in the accumulation of lipid peroxidation via the SLC7A11-GPX4 axis. Mechanistically, LncFASA directly binds to the Ahpc-TSA domain of PRDX1, inhibiting its peroxidase activity by driving liquid-liquid phase separation, which disrupts intracellular ROS homeostasis. Notably, high LncFASA expression indicates favorable overall survival in individuals with breast cancer, and LncFASA impairs the growth of breast xenograft tumors by modulating ferroptosis. Together, our findings illustrate the crucial role of this lncRNA in ferroptosis-mediated cancer development and provide new insights into therapeutic strategies for breast cancer.

LncRNA LINK-A Remodels Tissue Inflammatory Microenvironments to Promote Obesity.

High-fat diet (HFD)-induced obesity is a crucial risk factor for metabolic syndrome, mainly due to adipose tissue dysfunctions associated with it. However, the underlying mechanism remains unclear. This study has used genetic screening to identify an obesity-associated human lncRNA LINK-A as a critical molecule bridging the metabolic microenvironment and energy expenditure in vivo by establishing the HFD-induced obesity knock-in (KI) mouse model. Mechanistically, HFD LINK-A KI mice induce the infiltration of inflammatory factors, including IL-1ß and CXCL16, through the LINK-A/HB-EGF/HIF1α feedback loop axis in a self-amplified manner, thereby promoting the adipose tissue microenvironment remodeling and adaptive thermogenesis disorder, ultimately leading to obesity and insulin resistance. Notably, LINK-A expression is positively correlated with inflammatory factor expression in individuals who are overweight. Of note, targeting LINK-A via nucleic acid drug antisense oligonucleotides (ASO) attenuate HFD-induced obesity and metabolic syndrome, pointing out LINK-A as a valuable and effective therapeutic target for treating HFD-induced obesity. Briefly, the results reveale the roles of lncRNAs (such as LINK-A) in remodeling tissue inflammatory microenvironments to promote HFD-induced obesity.

Comparison of baseline drifts using three reflector blocks versus using a single reflector block for the calibration of wall-mounted Respiratory Gating for Scanner (RGSC) camera integrated with a CT.

BACKGROUND: The calibration of the Respiratory Gating for SCanner (RGSC) system is critical to achieve better and more stable accuracy. The current procedure for a wall-mounted RGSC system has a relatively large residual error. PURPOSE: To compare the baseline drifts in the image acquisition of DIBH using three reflector blocks versus using a single reflector block in the calibration of a wall-mounted RGSC camera system. MATERIALS AND METHODS: Varian provides a calibration plate with three rows of calibration points: each row is separated by 15 cm longitudinally and by 10 cm laterally. In Varian's single-block calibration method, the reflector block was first placed on the center point of the calibration plate and aligned with the scanner isocenter. The calibration took a picture of the block, then placed the block on the other eight points sequentially. In the proposed three-block method, we placed three reflector blocks on the center row, with the center block aligned with the isocenter, and we took a picture of the center block by manually blocking the other two blocks in calibration. By moving the couch longitudinally in or out 15 cm, the calibration goes through all nine points. Monte Carlo simulation was done using Matlab to analyze the calibration matrix eigenvalue characteristics. RESULTS: For a typical scan length of 40 cm of DIBH, the residual baseline drift in simulated DIBH is 0.02 ± 0.03  versus 0.30 ± 0.12 cm for three-block calibration and single-block calibration, respectively. To achieve 0.5 mm tolerance for the eigenvalue, the laser and reflector box should be within ±3 mm uncertainties based on the eigenvalue simulation. CONCLUSION: Three-block calibration method effectively removes baseline drift caused by couch movement in DIBH/4D CT scan for the wall-mounted camera while the single-block calibration method still has significant residual baseline drift.

AgingReG: a curated database of aging regulatory relationships in humans.

Aging and cellular senescence are characterized by a progressive loss of physiological integrity, which could be triggered by aging factors such as physiological, pathological and external factors. Numerous studies have shown that gene regulatory events play crucial roles in aging, increasing the need for a comprehensive repository of regulatory relationships during aging. Here, we established a manually curated database of aging factors (AgingReG, https://bio.liclab.net/Aging-ReG/), focusing on the regulatory relationships during aging with experimental evidence in humans. By curating thousands of published literature, 2157 aging factor entries (1345 aging gene entries, 804 external factor entries and eight aging-related pathway entries) and related regulatory information were manually curated. The regulatory relationships were classified into four types according to their functions: (i) upregulation, which indicates that aging factors upregulate the expression of target genes during aging; (ii) downregulation, which indicates that aging factors downregulate the expression of target genes during aging; (iii) activation, which indicates that aging factors influence the activity of target genes during aging and (iv) inhibition, which indicates that aging factors inhibit the activation of target molecule activity, leading to declined or lost target activity. AgingReG involves 651 upregulating pairs, 632 downregulating pairs, 330 activation-regulating pairs and 34 inhibition-regulating pairs, covering 195 disease types and more than 800 kinds of cells and tissues from 1784 published literature studies. AgingReG provides a user-friendly interface to query, browse and visualize detailed information about the regulatory relationships during aging. We believe that AgingReG will serve as a valuable resource database in the field of aging research. Database URL: https://bio.liclab.net/Aging-ReG/.

Clinical prospects and research strategies of long non-coding RNA encoding micropeptides. / é•¿é“¾éžç¼–ç RNAç¼–ç å°è‚½çš„ç”Ÿç‰©åŒ»å­¦æ„ä¹‰åŠç ”ç©¶ç­–ç•¥.

Long non-coding RNAs (lncRNAs) which are usually thought to have no protein coding ability, are widely involved in cell proliferation, signal transduction and other biological activities. However, recent studies have suggested that short open reading frames (sORFs) of some lncRNAs can encode small functional peptides (micropeptides). These micropeptides appear to play important roles in calcium homeostasis, embryonic development and tumorigenesis, suggesting their potential as therapeutic targets and diagnostic biomarkers. Currently, bioinformatic tools as well as experimental methods such as ribosome mapping and in vitro translation are applied to predict the coding potential of lncRNAs. Furthermore, mass spectrometry, specific antibodies and epitope tags are used for validating the expression of micropeptides. Here, we review the physiological and pathological functions of recently identified micropeptides as well as research strategies for predicting the coding potential of lncRNAs to facilitate the further research of lncRNA encoded micropeptides.

Real-time respiratory motion prediction using photonic reservoir computing.

Respiration induced motion is a well-recognized challenge in many clinical practices including upper body imaging, lung tumor motion tracking and radiation therapy. In this work, we present a recurrent neural network algorithm that was implemented in a photonic delay-line reservoir computer (RC) for real-time respiratory motion prediction. The respiratory motion signals are quasi-periodic waveforms subject to a variety of non-linear distortions. In this work, we demonstrated for the first time that RC can be effective in predicting short to medium range of respiratory motions within practical timescales. A double-sliding window technology is explored to enable the real-time establishment of an individually trained model for each patient and the real-time processing of live-streamed respiratory motion data. A breathing dataset from a total of 76 patients with breathing speeds ranging from 3 to 20 breaths per minute (BPM) is studied. Motion prediction of look-ahead times of 66.6, 166.6, and 333 ms are investigated. With a 333 ms look-ahead time, the real-time RC model achieves an average normalized mean square error (NMSE) of 0.025, an average mean absolute error (MAE) of 0.34 mm, an average root mean square error (RMSE) of 0.45 mm, an average therapeutic beam efficiency (TBE) of 94.14% for an absolute error (AE) < 1 mm, and 99.89% for AE < 3 mm. This study demonstrates that real-time RC is an efficient computing framework for high precision respiratory motion prediction.

LncRNA modulates Hippo-YAP signaling to reprogram iron metabolism.

Iron metabolism dysregulation is tightly associated with cancer development. But the underlying mechanisms remain poorly understood. Increasing evidence has shown that long noncoding RNAs (lncRNAs) participate in various metabolic processes via integrating signaling pathway. In this study, we revealed one iron-triggered lncRNA, one target of YAP, LncRIM (LncRNA Related to Iron Metabolism, also named ZBED5-AS1 and Loc729013), which effectively links the Hippo pathway to iron metabolism and is largely independent on IRP2. Mechanically, LncRIM directly binds NF2 to inhibit NF2-LATS1 interaction, which causes YAP activation and increases intracellular iron level via DMT1 and TFR1. Additionally, LncRIM-NF2 axis mediates cellular iron metabolism dependent on the Hippo pathway. Clinically, high expression of LncRIM correlates with poor patient survival, suggesting its potential use as a biomarker and therapeutic target. Taken together, our study demonstrated a novel mechanism in which LncRIM-NF2 axis facilitates iron-mediated feedback loop to hyperactivate YAP and promote breast cancer development.

Evaluation of repeatability and reproducibility of radiomic features produced by the fan-beam kV-CT on a novel ring gantry-based PET/CT linear accelerator.

BACKGROUND: The RefleXion X1 is a novel radiotherapy delivery system on a ring gantry equipped with fan-beam kV-CT and PET imaging subsystems. The day-to-day scanning variability of radiomics features must be evaluated before any attempt to utilize radiomics features. PURPOSE: This study aims to characterize the repeatability and reproducibility of radiomic features produced by the RefleXion X1 kV-CT. MATERIALS AND METHODS: The Credence Cartridge Radiomics (CCR) phantom includes six cartridges of varied materials. It was scanned 10 times on the RefleXion X1 kVCT imaging subsystem over a 3-month period using the two most frequently used scanning protocols (BMS and BMF). Fifty-five radiomic features were extracted for each ROI on each CT scan and analyzed using LifeX software. The coefficient of variation (COV) was computed to evaluate the repeatability. Intraclass correlation coefficient (ICC) and concordance correlation coefficient (CCC) were used to evaluate the repeatability and reproducibility of the scanned images using 0.9 as the threshold. This process is repeated on a GE PET-CT scanner using several built-in protocols as a comparison. RESULTS: On average, 87% of the features on both scan protocols on the RefleXion X1 kVCT imaging subsystem can be considered repeatable as they met COV < 10% criteria. On GE PET-CT, this number is similar at 86%. When we tighten the criteria to COV <5%, the RefleXion X1 kVCT imaging subsystem showed much better repeatability with 81% of features on average whereas GE PET-CT showed only 73.5% on average. About 91% and 89% of the features with ICC > 0.9 respectively for BMS and BMF protocols on RefleXion X1. On the other hand, the percentage of features with ICC > 0.9 on GE PET-CT ranges from 67% to 82%. The RefleXion X1 kVCT imaging subsystem showed excellent intra-scanner reproducibility between the scanning protocols much better than the GE PET CT scanner. For the inter-scanner reproducibility, the percentage of features with CCC > 0.9 ranged from 49% to 80%. between X1 and GE PET-CT scanning protocols. CONCLUSIONS: Clinically useful CT radiomic features produced by the RefleXion X1 kVCT imaging subsystem are reproducible and stable over time, demonstrating its utility as a quantitative imaging platform.

lncRNA BREA2 promotes metastasis by disrupting the WWP2-mediated ubiquitination of Notch1.

Notch has been implicated in human cancers and is a putative therapeutic target. However, the regulation of Notch activation in the nucleus remains largely uncharacterized. Therefore, characterizing the detailed mechanisms governing Notch degradation will identify attractive strategies for treating Notch-activated cancers. Here, we report that the long noncoding RNA (lncRNA) BREA2 drives breast cancer metastasis by stabilizing the Notch1 intracellular domain (NICD1). Moreover, we reveal WW domain containing E3 ubiquitin protein ligase 2 (WWP2) as an E3 ligase for NICD1 at K1821 and a suppressor of breast cancer metastasis. Mechanistically, BREA2 impairs WWP2-NICD1 complex formation and in turn stabilizes NICD1, leading to Notch signaling activation and lung metastasis. BREA2 loss sensitizes breast cancer cells to inhibition of Notch signaling and suppresses the growth of breast cancer patient-derived xenograft tumors, highlighting its therapeutic potential in breast cancer. Taken together, these results reveal the lncRNA BREA2 as a putative regulator of Notch signaling and an oncogenic player driving breast cancer metastasis.

MR-KPA: medication recommendation by combining knowledge-enhanced pre-training with a deep adversarial network.

BACKGROUND: Medication recommendation based on electronic medical record (EMR) is a research hot spot in smart healthcare. For developing computational medication recommendation methods based on EMR, an important challenge is the lack of a large number of longitudinal EMR data with time correlation. Faced with this challenge, this paper proposes a new EMR-based medication recommendation model called MR-KPA, which combines knowledge-enhanced pre-training with the deep adversarial network to improve medication recommendation from both feature representation and the fine-tuning process. Firstly, a knowledge-enhanced pre-training visit model is proposed to realize domain knowledge-based external feature fusion and pre-training-based internal feature mining for improving the feature representation. Secondly, a medication recommendation model based on the deep adversarial network is developed to optimize the fine-tuning process of pre-training visit model and alleviate over-fitting of model caused by the task gap between pre-training and recommendation. RESULT: The experimental results on EMRs from medical and health institutions in Hainan Province, China show that the proposed MR-KPA model can effectively improve the accuracy of medication recommendation on small-scale longitudinal EMR data compared with existing representative methods. CONCLUSION: The advantages of the proposed MR-KPA are mainly attributed to knowledge enhancement based on ontology embedding, the pre-training visit model and adversarial training. Each of these three optimizations is very effective for improving the capability of medication recommendation on small-scale longitudinal EMR data, and the pre-training visit model has the most significant improvement effect. These three optimizations are also complementary, and their integration makes the proposed MR-KPA model achieve the best recommendation effect.

Insight into a Bentonite-Based Hydrogel for the Conservation of Sandstone-Based Cultural Heritage: In Situ Formation, Reinforcement Mechanism, and High-Durability Evaluation.

Conservation of sandstone-based cultural heritage has attracted a great deal of interest. We propose herein a novel protecting strategy, via in situ fabrication of bentonite-based hydrogels (B-H) inside sandstones, where the bentonite-based hydrogels serve as the underlying cement. To create bentonite-based hydrogels with controllable structure, possessing good mechanical and anti-swelling properties, we have optimized forming time, appearance, and viscosity. The hydrogel precursor penetrated into the pores of the sandstone; the hydrogel would then form within 3-5 h. As found by employing a fluorescent tracer, the precursor remained controllably in place without any apparent change in the sandstone morphology. The bentonite-based hydrogels that formed inside the sandstones presented strong hydrogen bonding, coordination, and ionic bonding, as well as strong mechanical interlocking to the sandstone matrix. As a result, the sandstones possessed enhanced mechanical compressive strength and excellent resistance to acid, salt, and freeze-thaw cycles. Our approach provides for a non-destructive, eco-friendly, easy-to-use, and long-term strategy for cultural preservation, one with excellent protection effects.

Promoting anti-tumor immunity by targeting TMUB1 to modulate PD-L1 polyubiquitination and glycosylation.

Immune checkpoint blockade therapies targeting the PD-L1/PD-1 axis have demonstrated clear clinical benefits. Improved understanding of the underlying regulatory mechanisms might contribute new insights into immunotherapy. Here, we identify transmembrane and ubiquitin-like domain-containing protein 1 (TMUB1) as a modulator of PD-L1 post-translational modifications in tumor cells. Mechanistically, TMUB1 competes with HECT, UBA and WWE domain-containing protein 1 (HUWE1), a E3 ubiquitin ligase, to interact with PD-L1 and inhibit its polyubiquitination at K281 in the endoplasmic reticulum. Moreover, TMUB1 enhances PD-L1 N-glycosylation and stability by recruiting STT3A, thereby promoting PD-L1 maturation and tumor immune evasion. TMUB1 protein levels correlate with PD-L1 expression in human tumor tissue, with high expression being associated with poor patient survival rates. A synthetic peptide engineered to compete with TMUB1 significantly promotes antitumor immunity and suppresses tumor growth in mice. These findings identify TMUB1 as a promising immunotherapeutic target.

Exploring the Inhibitory Effect of RASFF on China-EU Trade of Rice-Based Products.

Rice-based products exported from China to Europe have repeatedly encountered technical trade barriers. Using panel data from 24 states of the European Union during 2001-2017, this study builds a theoretical model to investigate the impact of implementation, intensity and structure of the Rapid Alert System for Food and Feed (RASFF) on China-EU rice-based product trade. The study found that RASFF has a serious inhibitory effect on the trade of traditional rice-based products because of detecting GM ingredients, showing an obvious lag effect, diffusion effect and structure effect. The negative effect occurs in entry process, and the inhibitory effect of border rejection and information notifications results in time lag, but the marginal effect of alerts for market links is insignificant. Moreover, the positive information disclosure effect of technical barriers implemented by individual members is much smaller than the negative diffusion effect. Finally, countermeasures and suggestions are proposed, including the source supervision of the test, the supervision of GM variety approval and GM seed production, the establishment of an early-warning and rapid-response mechanism to technical barriers of agricultural products, and food enterprise information.

Tumor immune checkpoints and their associated inhibitors.

Immunological evasion is one of the defining characteristics of cancers, as the immune modification of an immune checkpoint (IC) confers immune evasion capabilities to tumor cells. Multiple ICs, such as programmed cell death protein-1 (PD-1) and cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4), can bind to their respective receptors and reduce tumor immunity in a variety of ways, including blocking immune cell activation signals. IC blockade (ICB) therapies targeting these checkpoint molecules have demonstrated significant clinical benefits. This is because antibody-based IC inhibitors and a variety of specific small molecule inhibitors can inhibit key oncogenic signaling pathways and induce durable tumor remission in patients with a variety of cancers. Deciphering the roles and regulatory mechanisms of these IC molecules will provide crucial theoretical guidance for clinical treatment. In this review, we summarize the current knowledge on the functional and regulatory mechanisms of these IC molecules at multiple levels, including epigenetic regulation, transcriptional regulation, and post-translational modifications. In addition, we provide a summary of the medications targeting various nodes in the regulatory pathway, and highlight the potential of newly identified IC molecules, focusing on their potential implications for cancer diagnostics and immunotherapy.

GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery.

BACKGROUND: Accurately predicting drug-target binding affinity (DTA) in silico plays an important role in drug discovery. Most of the computational methods developed for predicting DTA use machine learning models, especially deep neural networks, and depend on large-scale labelled data. However, it is difficult to learn enough feature representation from tens of millions of compounds and hundreds of thousands of proteins only based on relatively limited labelled drug-target data. There are a large number of unknown drugs, which never appear in the labelled drug-target data. This is a kind of out-of-distribution problems in bio-medicine. Some recent studies adopted self-supervised pre-training tasks to learn structural information of amino acid sequences for enhancing the feature representation of proteins. However, the task gap between pre-training and DTA prediction brings the catastrophic forgetting problem, which hinders the full application of feature representation in DTA prediction and seriously affects the generalization capability of models for unknown drug discovery. RESULTS: To address these problems, we propose the GeneralizedDTA, which is a new DTA prediction model oriented to unknown drug discovery, by combining pre-training and multi-task learning. We introduce self-supervised protein and drug pre-training tasks to learn richer structural information from amino acid sequences of proteins and molecular graphs of drug compounds, in order to alleviate the problem of high variance caused by encoding based on deep neural networks and accelerate the convergence of prediction model on small-scale labelled data. We also develop a multi-task learning framework with a dual adaptation mechanism to narrow the task gap between pre-training and prediction for preventing overfitting and improving the generalization capability of DTA prediction model on unknown drug discovery. To validate the effectiveness of our model, we construct an unknown drug data set to simulate the scenario of unknown drug discovery. Compared with existing DTA prediction models, the experimental results show that our model has the higher generalization capability in the DTA prediction of unknown drugs. CONCLUSIONS: The advantages of our model are mainly attributed to two kinds of pre-training tasks and the multi-task learning framework, which can learn richer structural information of proteins and drugs from large-scale unlabeled data, and then effectively integrate it into the downstream prediction task for obtaining a high-quality DTA prediction in unknown drug discovery.