Resveratrol modulates the Nrf2/NF-κB pathway and inhibits TSLP-mediated atopic march.

BACKGROUND: Resveratrol has been found to have anti-inflammatory and anti-allergic properties. The effects of resveratrol on thymic stromal lymphopoietin (TSLP)-mediated atopic march remain unclear. PURPOSE: To explore the potential role of resveratrol in TSLP-mediated atopic march. METHODS: The atopic march mouse model was established by topical application of MC903 (a vitamin D3 analog). Following the treatment with resveratrol, airway resistance in mice was discovered by pulmonary function apparatus, and the number of total cells, neutrophils, and eosinophils in bronchoalveolar lavage fluid was counted. The histopathological features of pulmonary and ear skin tissues, inflammation, and cell infiltration were determined by hematoxylin and eosin staining. The messenger RNA (mRNA) levels of TSLP, immunoglobulin E, interleukin (IL)-4, IL-5, and IL-13 were measured by real-time quantitative polymerase chain reaction. The protein expression of nuclear factor kappa B (NF-κB)/nuclear factor erythroid 2-related factor 2 (Nrf2) signaling-associated molecules (p-p65, p65, p-I kappa B kinase alpha (IκBα), IκBα, Nrf2, and TSLP) in lung and ear skin tissues were assessed by Western blot analysis. RESULTS: Resveratrol attenuated airway resistance and infiltration of total cells, eosinophils, and neutrophils in both lung and ear skin tissues. Resveratrol ameliorates serum inflammatory markers in allergic mice. Moreover, the phosphorylation levels of NF-κB pathway-related proteins were significantly reduced by administration of resveratrol in allergic lung and ear skin tissues. Similarly, the protein expression of TSLP in both lung and ear skin tissues was reduced by resveratrol, and Nrf2, a protector molecule, was increased with resveratrol treatment. CONCLUSION: Resveratrol attenuates TSLP-reduced atopic march through ameliorating inflammation and cell infiltration in pulmonary and ear skin tissues by inhibiting the abnormal activation of NF-κB signaling pathway.

BiMPADR: A Deep Learning Framework for Predicting Adverse Drug Reactions in New Drugs.

Detecting the unintended adverse reactions of drugs (ADRs) is a crucial concern in pharmacological research. The experimental validation of drug-ADR associations often entails expensive and time-consuming investigations. Thus, a computational model to predict ADRs from known associations is essential for enhanced efficiency and cost-effectiveness. Here, we propose BiMPADR, a novel model that integrates drug gene expression into adverse reaction features using a message passing neural network on a bipartite graph of drugs and adverse reactions, leveraging publicly available data. By combining the computed adverse reaction features with the structural fingerprints of drugs, we predict the association between drugs and adverse reactions. Our models obtained high AUC (area under the receiver operating characteristic curve) values ranging from 0.861 to 0.907 in an external drug validation dataset under differential experiment conditions. The case study on multiple BET inhibitors also demonstrated the high accuracy of our predictions, and our model's exploration of potential adverse reactions for HWD-870 has contributed to its research and development for market approval. In summary, our method would provide a promising tool for ADR prediction and drug safety assessment in drug discovery and development.

WaveICA 2.0: a novel batch effect removal method for untargeted metabolomics data without using batch information.

INTRODUCTION: Untargeted metabolomics based on liquid chromatography-mass spectrometry is inevitably affected by batch effects that are caused by non-biological systematic bias. Previously, we developed a novel method called WaveICA to remove batch effects for untargeted metabolomics data. To detect batch effect information, the method relies on a batch label. However, it cannot be used in the scenario in which there is only one batch of data or the batch label is unknown. OBJECTIVES: We aim to improve the WaveICA method to remove batch effects for untargeted metabolomics data without using batch information. METHODS: We improved the WaveICA method by developing WaveICA 2.0 to remove batch effects for metabolomics data, and provided an R package WaveICA_2.0 to implement this method. RESULTS: The performance of the WaveICA 2.0 method was evaluated on real metabolomics data. For metabolomics data with three batches, the performance of the WaveICA 2.0 method was similar to that of the WaveICA method in terms of gathering quality control samples (QCSs) and subject samples together in principle component analysis score plots, increasing the similarity of QCSs, increasing differential peaks, and improving classification accuracy. For metabolomics data with only one batch, the WaveICA 2.0 method had a strong ability to remove intensity drift and reveal more biological information and outperformed the QC-RLSC and QC-SVRC methods in our study using our metabolomics data. CONCLUSION: Our results demonstrated that the WaveICA 2.0 method can be used in practice to remove batch effects for untargeted metabolomics data without batch information.

Distinct associations of NEDD4L expression with genetic abnormalities and prognosis in acute myeloid leukemia.

BACKGROUND: There is mounting evidence that demonstrated the association of aberrant NEDD4L expression with diverse human cancers. However, the expression pattern and clinical implication of NEDD4L in acute myeloid leukemia (AML) remains poorly defined. METHODS: We systemically determined NEDD4L expression with its clinical significance in AML by both public data and our research cohort. Moreover, biological functions of NEDD4L in leukemogenesis were further tested by in vitro experiments. RESULTS: By the public data, we identified that low NEDD4L expression was correlated with AML among diverse human cancers. Expression of NEDD4L was remarkably decreased in AML compared with controls, and was confirmed by our research cohort. Clinically, low expression of NEDD4L was correlated with greatly lower age, higher white blood cells, and higher bone marrow/peripheral blood blasts. Moreover, NEDD4L underexpression was positively correlated with normal karyotype, FLT3 and NPM1 mutations, but negatively associated with complex karyotype and TP53 mutations. Importantly, the association between NEDD4L expression and survival was also discovered in cytogenetically normal AML patients. Finally, a number of 1024 RNAs and 91 microRNAs were identified to be linked to NEDD4L expression in AML. Among the negatively correlated microRNAs, miR-10a was also discovered as a microRNA that may directly target NEDD4L. Further functional studies revealed that NEDD4L exhibited anti-proliferative and pro-apoptotic effects in leukemic cell line K562. CONCLUSIONS: Our findings indicated that NEDD4L underexpression, as a frequent event in AML, was associated with genetic abnormalities and prognosis in AML. Moreover, NEDD4L expression may be involved in leukemogenesis with potential therapeutic target value.

Incidence of and risk factors associated with lung metastases in newly diagnosed epithelial ovarian cancer with a look on prognosis after diagnosis: a population-based cohort study of the SEER database.

PURPOSE: Patients with lung metastases (LM) from epithelial ovarian cancer (EOC) (EOCLM) usually have a poor prognosis. However, there is no consensus on the optimal management of these patients. In this study, we aimed to take a look at the incidence of LM and factors associated with its occurrence as well as the prognosis in newly diagnosed EOC with LM on a population level. METHODS: EOC patients diagnosed between the years 2010 and 2016 were identified from the Surveillance, Epidemiology, and End Results (SEER) program database. Multivariable logistic regression and multivariable Cox regression were used to investigate the factors that could predict the occurrence of and prognosis after diagnosis of EOC with LM. RESULTS: Of the 33,418 qualified EOC patients, 2240 (6.7%) were noted to have LMs at the time of EOC diagnosis. Higher T stage, N1 stage, advanced tumor grade, and elevated cancer antigen-125 levels were found to be associated with a higher risk of having LM at the time of EOC diagnosis. The median survival time after diagnosis with EOCLM was found to be 13.0 months (interquartile range: 3.0-34.0 months). Being unmarried and having mucinous histology were both associated with increased all-cause death risk from EOCLM. However, the primary tumor originated from the midline of ovaries, surgical management, and whether patient received chemotherapy or not predicted improved overall survival. The median survival time of patients was significantly longer for EOCLM cases managed surgically (31.0 months) versus those who did not have surgery (4.0 months), as well as EOCLM cases received chemotherapy (23.0 months) versus those who did not have chemotherapy (2.0 months). CONCLUSION: This retrospective cohort study showed that de novo LM was infrequent in EOC patients overall and when present predicted poor prognosis. The findings can be potentially useful in formulating for follow-up strategies, screening tools, and personalized interventions.

NormAE: Deep Adversarial Learning Model to Remove Batch Effects in Liquid Chromatography Mass Spectrometry-Based Metabolomics Data.

Untargeted metabolomics based on liquid chromatography-mass spectrometry is affected by nonlinear batch effects, which cover up biological effects, result in nonreproducibility, and are difficult to be calibrate. In this study, we propose a novel deep learning model, called Normalization Autoencoder (NormAE), which is based on nonlinear autoencoders (AEs) and adversarial learning. An additional classifier and ranker are trained to provide adversarial regularization during the training of the AE model, latent representations are extracted by the encoder, and then the decoder reconstructs the data without batch effects. The NormAE method was tested on two real metabolomics data sets. After calibration by NormAE, the quality control samples (QCs) for both data sets gathered most closely in a PCA score plot (average distances decreased from 56.550 and 52.476 to 7.383 and 14.075, respectively) and obtained the highest average correlation coefficients (from 0.873 and 0.907 to 0.997 for both). Additionally, NormAE significantly improved biomarker discovery (median number of differential peaks increased from 322 and 466 to 1140 and 1622, respectively). NormAE was compared with four commonly used batch effect removal methods. The results demonstrated that using NormAE produces the best calibration results.

Three-wavelength passive demodulation technique for the interrogation of EFPI sensors with arbitrary cavity length.

A three-wavelength passive demodulation technique to interrogate extrinsic Fabry-Perot interferometric (EFPI) sensors with arbitrary cavity length has been developed. The DC component is obtained online, then the applied dynamic signal is recovered by using three new signals without the DC component. The performance of the technique is demonstrated by simulations and experiments. The demodulation technique can extract dynamic signals, regardless of whether the phase modulation is larger than 2π. Theoretically, EFPI sensors with arbitrary cavity length can be demodulated by the demodulation technique, and EFPI sensors with cavity lengths in the 22.96-1002.3 µm range are detected successfully by the same demodulator in experiments. The technique is robust with respect to the bending loss of the leading fiber. The demodulation technique provides a robust and accurate solution to measure dynamic signals for EFPI sensors. It has the properties of high frequency, a large dynamic range, and high sensitivity. The paper demonstrates this technique's potential for measuring dynamic signals.

Antiresonant mechanism based self-temperature-calibrated fiber optic Fabry-Perot gas pressure sensors.

A self-temperature-calibrated gas pressure sensor with a sandwich structure made of single-mode fiber (SMF)-hollow core fiber (HCF)-SMF is proposed and experimentally demonstrated. A Fabry-Perot interferometer (FPI) is formed by the SMF-HCF-SMF structure along the axial direction, and an antiresonant reflecting optical waveguide (ARROW) is formed by the ring-cladding of the HCF along the radial direction. A micro-channel is drilled on the ring-cladding of the HCF using a femtosecond laser to facilitate air entering/exiting the HCF. The FPI functions as the pressure sensor, and the ARROW functions as the temperature sensor. The initial wavelength and pressure sensitivity of the FPI can be calibrated from the temperature obtained by measuring the optical thickness of the ARROW. The experimental results show that the ARROW exhibits a temperature sensitivity of ~0.584 nm/°C, and the pressure sensitivity of the FPI ranges from 3.884 to 0.919 nm/MPa, within the temperature range of 37-1007 °C. The simplicity and durability of the sensor make it suitable for reliable gas pressure measurement in high-temperature environments.

Dysregulation of miR-200s clusters as potential prognostic biomarkers in acute myeloid leukemia.

BACKGROUND: Increasing studies showed that miR-200 family (miR-200s) clusters are aberrantly expressed in multiple human cancers, and miR-200s clusters function as tumor suppressor genes by affecting cell proliferation, self-renewal, differentiation, division and apoptosis. Herein, we aimed to investigate the expression and clinical implication of miR-200s clusters in acute myeloid leukemia (AML). METHODS: RT-qPCR was performed to detect expression of miR-200s clusters in 19 healthy donors, 98 newly diagnosed AML patients, and 35 AML patients achieved complete remission (CR). RESULTS: Expression of miR-200a/200b/429 cluster but not miR-200c/141 cluster was decreased in newly diagnosed AML patients as compared to healthy donors and AML patients achieved CR. Although no significant differences were observed between miR-200s clusters and most of the features, low expression of miR-200s clusters seems to be associated with higher white blood cells especially for miR-200a/200b. Of the five members of miR-200s clusters, low expression of miR-200b/429/200c was found to be associated with lower CR rate. Logistic regression analysis further revealed that low expression of miR-429 acted as an independent risk factor for CR in AML. Based on Kaplan-Meier analysis, low expression of miR-200b/429/200c was associated with shorter OS, whereas miR-200a/141 had a trend. Moreover, multivariate analysis of Cox regression models confirmed the independently prognostic value of miR-200b expression for OS in AML. CONCLUSIONS: Expression of miR-200a/200b/429 cluster was frequently down-regulated in AML, and low expression of miR-429 as an independent risk factor for CR, whereas low expression of miR-200b as an independent prognostic biomarker for OS.

Five-step phase-shifting white-light interferometry for the measurement of fiber optic extrinsic Fabry-Perot interferometers.

Five-step phase-shifting white-light interferometry is presented for interrogating the absolute cavity length of the fiber optic extrinsic Fabry-Perot interferometer (EFPI). It combines ideas of phase-shifting interferometry and white-light interferometry (WLI) to extend the measurement range of fiber optic WLI. Five sub-interferograms intercepted from the white-light optical spectrum are used to recover the optical path difference (OPD) of the EFPI. This method is demonstrated to interrogate a wider range of OPD. The experimental results show that the measurement resolution ranges from 0.5 µm to 5 µm with cavity length ranges from 16 µm to 12,402 µm, and it has a great advantage in measuring EFPIs with short cavity lengths.

Identifying survival of pan-cancer patients under immunotherapy using genomic mutation signature with large sample cohorts.

Although immune checkpoint inhibitors have led to durable clinical response in multiple cancers, only a small proportion of patients respond to this treatment. Therefore, we aim to develop a predictive model that utilizes gene mutation profiles to accurately identify the survival of pan-cancer patients with immunotherapy. Here, we develop and evaluate three different nomograms using two cohorts containing 1,594 cancer patients whose mutation profiles are obtained by MSK-IMPACT sequencing and 230 cancer patients receiving whole-exome sequencing, respectively. Using eighteen genes (SETD2, BRAF, NCOA3, LATS1, IL7R, CREBBP, TET1, EPHA7, KDM5C, MET, KMT2D, RET, PAK7, CSF1R, JAK2, FAT1, ASXL1 and SPEN), the first nomogram stratifies patients from both cohorts into High-Risk and Low-Risk groups. Pan-cancer patients in the High-Risk group exhibit significantly shorter overall survival and progression-free survival than patients in the Low-Risk group in both cohorts. Meanwhile, the first nomogram also accurately identifies the survival of patients with melanoma or lung cancer undergoing immunotherapy, or pan-cancer patients treated with anti-PD-1/PD-L1 inhibitor or anti-CTLA-4 inhibitor. The model proposed is not a prognostic model for the survival of pan-cancer patients without immunotherapy, but a simple, effective and robust predictive model for pan-cancer patients' survival under immunotherapy, and could provide valuable assistance for clinical practice.

A bidirectional interpretable compound-protein interaction prediction framework based on cross attention.

The identification of compound-protein interactions (CPIs) plays a vital role in drug discovery. However, the huge cost and labor-intensive nature in vitro and vivo experiments make it urgent for researchers to develop novel CPI prediction methods. Despite emerging deep learning methods have achieved promising performance in CPI prediction, they also face ongoing challenges: (i) providing bidirectional interpretability from both the chemical and biological perspective for the prediction results; (ii) comprehensively evaluating model generalization performance; (iii) demonstrating the practical applicability of these models. To overcome the challenges posed by current deep learning methods, we propose a cross multi-head attention oriented bidirectional interpretable CPI prediction model (CmhAttCPI). First, CmhAttCPI takes molecular graphs and protein sequences as inputs, utilizing the GCW module to learn atom features and the CNN module to learn residue features, respectively. Second, the model applies cross multi-head attention module to compute attention weights for atoms and residues. Finally, CmhAttCPI employs a fully connected neural network to predict scores for CPIs. We evaluated the performance of CmhAttCPI on balanced datasets and imbalanced datasets. The results consistently show that CmhAttCPI outperforms multiple state-of-the-art methods. We constructed three scenarios based on compound and protein clustering and comprehensively evaluated the model generalization ability within these scenarios. The results demonstrate that the generalization ability of CmhAttCPI surpasses that of other models. Besides, the visualizations of attention weights reveal that CmhAttCPI provides chemical and biological interpretation for CPI prediction. Moreover, case studies confirm the practical applicability of CmhAttCPI in discovering anticancer candidates.

O-GlcNAcylation of melanophilin enhances radiation resistance in glioblastoma via suppressing TRIM21 mediated ubiquitination.

The molecular mechanism of glioblastoma (GBM) radiation resistance remains poorly understood. The aim of this study was to elucidate the potential role of Melanophilin (MLPH) O-GlcNAcylation and the specific mechanism through which it regulates GBM radiotherapy resistance. We found that MLPH was significantly upregulated in recurrent GBM tumor tissues after ionizing radiation (IR). MLPH induced radiotherapy resistance in GBM cells and xenotransplanted human tumors through regulating the NF-κB pathway. MLPH was O-GlcNAcylated at the conserved serine 510, and radiation-resistant GBM cells showed higher levels of O-GlcNAcylation of MLPH. O-GlcNAcylation of MLPH protected its protein stability and tripartite motif containing 21(TRIM21) was identified as an E3 ubiquitin ligase promoting MLPH degradation whose interaction with MLPH was affected by O-GlcNAcylation. Our data demonstrate that MLPH exerts regulatory functions in GBM radiation resistance by promoting the NF-κB signaling pathway and that O-GlcNAcylation of MLPH both stabilizes and protects it from TRIM21-mediated ubiquitination. These results identify a potential mechanism of GBM radiation resistance and suggest a potential therapeutic strategy for GBM treatment.

Atox1 protects hippocampal neurons after traumatic brain injury via DJ-1 mediated anti-oxidative stress and mitophagy.

Regulation of the oxidative stress response is crucial for the management and prognosis of traumatic brain injury (TBI). The copper chaperone Antioxidant 1 (Atox1) plays a crucial role in regulating intracellular copper ion balance and impacting the antioxidant capacity of mitochondria, as well as the oxidative stress state of cells. However, it remains unknown whether Atox1 is involved in modulating oxidative stress following TBI. Here, we investigated the regulatory role of Atox1 in oxidative stress on neurons both in vivo and in vitro, and elucidated the underlying mechanism through culturing hippocampal HT-22 cells with Atox1 mutation. The expression of Atox1 was significantly diminished following TBI, while mice with overexpressed Atox1 exhibited a more preserved hippocampal structure and reduced levels of oxidative stress post-TBI. Furthermore, the mice displayed notable impairments in learning and memory functions after TBI, which were ameliorated by the overexpression of Atox1. In the stretch injury model of HT-22 cells, overexpression of Atox1 mitigated oxidative stress by preserving the normal morphology and network connectivity of mitochondria, as well as facilitating the elimination of damaged mitochondria. Mechanistically, co-immunoprecipitation and mass spectrometry revealed the binding of Atox1 to DJ-1. Knockdown of DJ-1 in HT-22 cells significantly impaired the antioxidant capacity of Atox1. Mutations in the copper-binding motif or sequestration of free copper led to a substantial decrease in the interaction between Atox1 and DJ-1, with overexpression of DJ-1 failing to restore the antioxidant capacity of Atox1 mutants. The findings suggest that DJ-1 mediates the ability of Atox1 to withstand oxidative stress. And targeting Atox1 could be a potential therapeutic approach for addressing post-traumatic neurological dysfunction.

Meningioma achieves malignancy and erastin-induced ferroptosis resistance through FOXM1-AURKA-NRF2 axis.

The oncogene Aurora kinase A (AURKA) has been implicated in various tumor, yet its role in meningioma remains unexplored. Recent studies have suggested a potential link between AURKA and ferroptosis, although the underlying mechanisms are unclear. This study presented evidence of AURKA upregulation in high grade meningioma and its ability to enhance malignant characteristics. We identified AURKA as a suppressor of erastin-induced ferroptosis in meningioma. Mechanistically, AURKA directly interacted with and phosphorylated kelch-like ECH-associated protein 1 (KEAP1), thereby activating nuclear factor erythroid 2 related factor 2 (NFE2L2/NRF2) and target genes transcription. Additionally, forkhead box protein M1 (FOXM1) facilitated the transcription of AURKA. Suppression of AURKA, in conjunction with erastin, yields significant enhancements in the prognosis of a murine model of meningioma. Our study elucidates an unidentified mechanism by which AURKA governs ferroptosis, and strongly suggests that the combination of AURKA inhibition and ferroptosis-inducing agents could potentially provide therapeutic benefits for meningioma treatment.

A Novel Acetylation-Immune Subtyping for the Identification of a BET Inhibitor-Sensitive Subgroup in Melanoma.

BACKGROUND: There have been significant advancements in melanoma therapies. BET inhibitors (BETis) show promise in impairing melanoma growth. However, identifying BETi-sensitive melanoma subtypes is challenging. METHODS AND RESULTS: We analyzed 48 melanoma cell lines and 104 patients and identified two acetylation-immune subtypes (ALISs) in the cell lines and three ALISs in the patients. ALIS I, with high HAT1 and low KAT2A expression, showed a higher sensitivity to the BETi JQ-1 than ALIS II. ALIS III had low HAT1 expression. The TAD2B expression was low in ALIS I and II. KAT2A and HAT1 expressions were negatively correlated with the methylation levels of their CG sites (p = 0.0004 and 0.0003). Immunological gene sets, including B cell metagenes, activated stroma-related genes, fibroblast TGF response signatures (TBRS), and T cell TBRS-related genes, were up-regulated in ALIS I. Furthermore, KAT2A played a key role in regulating BETi sensitivity. CONCLUSIONS: The sensitivity of ALIS I to the BETi JQ-1 may be due to the inhibition of BETi resistance pathways and genes by low KAT2A expression and the dysregulation of the immune microenvironment by high HAT1 expression resulting from the absence of immune cells. ALIS I had the worst progression but showed sensitivity to BETi and B-cell-related immunotherapy, despite not responding to BRAF inhibitors.

Associations between the compositional patterns of blood volatile organic compounds and chronic respiratory diseases and ages at onset in NHANES 2003-2012.

BACKGROUND: and Purpose Volatile organic compounds (VOCs) pose a serious respiratory hazard. This study evaluated the relationship between the compositional patterns of blood VOCs and the risk and age at onset of chronic respiratory diseases (CRDs), including asthma, emphysema and chronic bronchitis, with the objective of preventing or delaying CRDs. METHODS: Participants from five cycles of the NHANES survey were included. Blood VOCs were clustered using k-means clustering. Differences in VOCs and age at onset between multiple groups were compared with the Kruskal‒Wallis test. Logistic regression and a generalized linear model were applied to examine the associations between different compositional patterns of blood VOCs and risk and age at onset of CRDs. RESULTS: 12,386 participants were enrolled in this study. Three VOC compositional patterns were identified after clustering nine species of blood VOCs. The concentration of VOCs in pattern 2 was relatively low and stable. The concentrations of benzene, ethylbenzene, o-xylene, styrene, toluene and m-p-xylene in pattern 3 and the concentrations of 1,4-dichlorobenzene and MTBE in pattern 1 were significantly higher than those in pattern 2. After adjustment for covariates, the participants with VOC pattern 3 had an increased risk of asthma (OR = 1.23, 95% CI: 1.02, 1.49), emphysema (OR = 3.37, 95% CI: 2.24, 5.06) and chronic bronchitis (OR = 1.79, 95% CI: 1.30, 2.45). Meanwhile, VOC pattern 3 was negatively correlated with the age at onset of asthma (ß = -5.61, 95% CI: 9.69, -1.52) and chronic bronchitis (ß = -9.17, 95% CI: 13.96, -4.39). VOC pattern 1 was not associated with either risk or age at onset of the three CRDs after adjustment. CONCLUSIONS: Changing the compositional pattern of blood VOCs by reducing certain species of VOCs may be a new strategy to lengthen the ages at onset of CRDs and effectively prevent them.

HiRAND: A novel GCN semi-supervised deep learning-based framework for classification and feature selection in drug research and development.

The prediction of response to drugs before initiating therapy based on transcriptome data is a major challenge. However, identifying effective drug response label data costs time and resources. Methods available often predict poorly and fail to identify robust biomarkers due to the curse of dimensionality: high dimensionality and low sample size. Therefore, this necessitates the development of predictive models to effectively predict the response to drugs using limited labeled data while being interpretable. In this study, we report a novel Hierarchical Graph Random Neural Networks (HiRAND) framework to predict the drug response using transcriptome data of few labeled data and additional unlabeled data. HiRAND completes the information integration of the gene graph and sample graph by graph convolutional network (GCN). The innovation of our model is leveraging data augmentation strategy to solve the dilemma of limited labeled data and using consistency regularization to optimize the prediction consistency of unlabeled data across different data augmentations. The results showed that HiRAND achieved better performance than competitive methods in various prediction scenarios, including both simulation data and multiple drug response data. We found that the prediction ability of HiRAND in the drug vorinostat showed the best results across all 62 drugs. In addition, HiRAND was interpreted to identify the key genes most important to vorinostat response, highlighting critical roles for ribosomal protein-related genes in the response to histone deacetylase inhibition. Our HiRAND could be utilized as an efficient framework for improving the drug response prediction performance using few labeled data.

Biologically Interpretable Deep Learning To Predict Response to Immunotherapy In Advanced Melanoma Using Mutations and Copy Number Variations.

Only 30-40% of advanced melanoma patients respond effectively to immunotherapy in clinical practice, so it is necessary to accurately identify the response of patients to immunotherapy pre-clinically. Here, we develop KP-NET, a deep learning model that is sparse on KEGG pathways, and combine it with transfer- learning to accurately predict the response of advanced melanomas to immunotherapy using KEGG pathway-level information enriched from gene mutation and copy number variation data. The KP-NET demonstrates best performance with AUROC of 0.886 on testing set and 0.803 on an unseen evaluation set when predicting responders (CR/PR/SD with PFS ≥6 mo) versus non-responders (PD/SD with PFS <6 mo) in anti-CTLA-4 treated melanoma patients. The model also achieves an AUROC of 0.917 and 0.833 in predicting CR/PR versus PD, respectively. Meanwhile, the AUROC is 0.913 when predicting responders versus non-responders in anti-PD-1/PD-L1 melanomas. Moreover, the KP-NET reveals some genes and pathways associated with response to anti-CTLA-4 treatment, such as genes PIK3CA, AOX1 and CBLB, and ErbB signaling pathway, T cell receptor signaling pathway, et al. In conclusion, the KP-NET can accurately predict the response of melanomas to immunotherapy and screen related biomarkers pre-clinically, which can contribute to precision medicine of melanoma.

Advances of Artificial Intelligence in Anti-Cancer Drug Design: A Review of the Past Decade.

Anti-cancer drug design has been acknowledged as a complicated, expensive, time-consuming, and challenging task. How to reduce the research costs and speed up the development process of anti-cancer drug designs has become a challenging and urgent question for the pharmaceutical industry. Computer-aided drug design methods have played a major role in the development of cancer treatments for over three decades. Recently, artificial intelligence has emerged as a powerful and promising technology for faster, cheaper, and more effective anti-cancer drug designs. This study is a narrative review that reviews a wide range of applications of artificial intelligence-based methods in anti-cancer drug design. We further clarify the fundamental principles of these methods, along with their advantages and disadvantages. Furthermore, we collate a large number of databases, including the omics database, the epigenomics database, the chemical compound database, and drug databases. Other researchers can consider them and adapt them to their own requirements.