Brain age prediction via cross-stratified ensemble learning.

As an important biomarker of neural aging, the brain age reflects the integrity and health of the human brain. Accurate prediction of brain age could help to understand the underlying mechanism of neural aging. In this study, a cross-stratified ensemble learning algorithm with staking strategy was proposed to obtain brain age and the derived predicted age difference (PAD) using T1-weighted magnetic resonance imaging (MRI) data. The approach was characterized as by implementing two modules: one was three base learners of 3D-DenseNet, 3D-ResNeXt, 3D-Inception-v4; another was 14 secondary learners of liner regressions. To evaluate performance, our method was compared with single base learners, regular ensemble learning algorithms, and state-of-the-art (SOTA) methods. The results demonstrated that our proposed model outperformed others models, with three metrics of mean absolute error (MAE), root mean-squared error (RMSE), and coefficient of determination (R2) of 2.9405 years, 3.9458 years, and 0.9597, respectively. Furthermore, there existed significant differences in PAD among the three groups of normal control (NC), mild cognitive impairment (MCI) and Alzheimer's disease (AD), with an increased trend across NC, MCI, and AD. It was concluded that the proposed algorithm could be effectively used in computing brain aging and PAD, and offering potential for early diagnosis and assessment of normal brain aging and AD.

Clinical application of ctDNA in early diagnosis, treatment and prognosis of patients with non-small cell lung cancer.

Lung cancer is one of the most common malignancies worldwide, with non-small cell lung cancer (NSCLC) being the most common type. As understanding of precise treatment options for NSCLC deepens, circulating tumor DNA (ctDNA) has emerged as a potential biomarker that has become a research hotspot and may represent a new approach for the individualized diagnosis and treatment of NSCLC. This article reviews the applications of ctDNA for the early screening of patients with NSCLC, guiding targeted therapy and immunotherapy, evaluating chemotherapy and postoperative efficacy, assessing prognosis and monitoring recurrence. With the in-depth study of the pathogenesis of NSCLC, plasma ctDNA may become an indispensable part of the precise treatment of NSCLC, which has great clinical application prospects.

[Box: see text].

Application of visual transformer in renal image analysis.

Deep Self-Attention Network (Transformer) is an encoder-decoder architectural model that excels in establishing long-distance dependencies and is first applied in natural language processing. Due to its complementary nature with the inductive bias of convolutional neural network (CNN), Transformer has been gradually applied to medical image processing, including kidney image processing. It has become a hot research topic in recent years. To further explore new ideas and directions in the field of renal image processing, this paper outlines the characteristics of the Transformer network model and summarizes the application of the Transformer-based model in renal image segmentation, classification, detection, electronic medical records, and decision-making systems, and compared with CNN-based renal image processing algorithm, analyzing the advantages and disadvantages of this technique in renal image processing. In addition, this paper gives an outlook on the development trend of Transformer in renal image processing, which provides a valuable reference for a lot of renal image analysis.

Transcription factor Nrf2 binds to circRNAPIBF1 to regulate SOD2 in lung adenocarcinoma progression.

Emerging evidence indicates that circular RNAs (circRNAs) play important roles in disease development, especially in cancers. Analysis of circRNA expression microarrays from the Gene Expression Omnibus database revealed that circPIBF1 was highly upregulated in lung adenocarcinoma (LUAD). The main aim of this study was to probe the function of circPIBF1 in pyroptosis of LUAD cells and the signal transduction pathways involved. CircPIBF1 was significantly overexpressed in LUAD and was related to the dismal prognosis of patients with LUAD. CircPIBF1 could bind to nuclear factor erythroid 2-related factor 2 (Nrf2), which further promoted the expression of superoxide dismutase 2 (SOD2). In addition, Nrf2 was also observed to recruit histone acetyltransferase E1A binding protein p300 (EP300) to enhance H3K27ac modification of SOD2, thus modulating the Nrf2-Keap1 signaling pathway. Moreover, we found that knockdown of circPIBF1 significantly suppressed the expression of SOD2 in cells and LUAD cell growth, while enhanced the expression of pyroptosis-related factors, which were further reversed by overexpression of SOD2 or EP300. Collectively, our findings suggest a direct involvement of circPIBF1 in pyroptosis-related LUAD carcinogenesis and implicate a role of Nrf2/EP300/SOD2 signaling in this process.

Resveratrol against Cardiac Fibrosis: Research Progress in Experimental Animal Models.

Cardiac fibrosis is a heterogeneous disease, which is characterized by abundant proliferation of interstitial collagen, disordered arrangement, collagen network reconstruction, increased cardiac stiffness, and decreased systolic and diastolic functions, consequently developing into cardiac insufficiency. With several factors participating in and regulating the occurrence and development of cardiac fibrosis, a complex molecular mechanism underlies the disease. Moreover, cardiac fibrosis is closely related to hypertension, myocardial infarction, viral myocarditis, atherosclerosis, and diabetes, which can lead to serious complications such as heart failure, arrhythmia, and sudden cardiac death, thus seriously threatening human life and health. Resveratrol, with the chemical name 3,5,4'-trihydroxy-trans-stilbene, is a polyphenol abundantly present in grapes and red wine. It is known to prevent the occurrence and development of cardiovascular diseases. In addition, it may resist cardiac fibrosis through a variety of growth factors, cytokines, and several cell signaling pathways, thus exerting a protective effect on the heart.

Retrograde Cerebral Perfusion Results in Better Perfusion to the Striatum Than the Cerebral Cortex During Deep Hypothermic Circulatory Arrest: A Microdialysis Study.

It remains controversial whether contemporary cerebral perfusion techniques, utilized during deep hypothermic circulatory arrest (DHCA), establish adequate perfusion to deep structures in the brain. This study aimed to investigate whether selective antegrade cerebral perfusion (SACP) or retrograde cerebral perfusion (RCP) can provide perfusion equally to various anatomical positions in the brain using metabolic evidence obtained from microdialysis. Eighteen piglets were randomly assigned to 40 min of circulatory arrest (CA) at 18°C without cerebral perfusion (DHCA group, n = 6) or with SACP (SACP group, n = 6) or RCP (RCP group, n = 6). Microdialysis parameters (glucose, lactate, pyruvate, and glutamate) were measured every 30 min in cortex and striatum. After 3 h of reperfusion, brain tissue was harvested for Western blot measurement of α-spectrin. After 40 min of CA, the DHCA group showed marked elevations of lactate and glycerol and a reduction in glucose in the microdialysis perfusate (all P < 0.05). The changes in glucose, lactate, and glycerol in the perfusate and α-spectrin expression in brain tissue were similar between cortex and striatum in the SACP group (all P > 0.05). In the RCP group, the cortex exhibited lower glucose, higher lactate, and higher glycerol in the perfusate and higher α-spectrin expression in brain tissue compared with the striatum (all P < 0.05). Glutamate showed no difference between cortex and striatum in all groups (all P > 0.05). In summary, SACP provided uniform and continuous cerebral perfusion to most anatomical sites in the brain, whereas RCP resulted in less sufficient perfusion to the cortex but better perfusion to the striatum.

Is selective antegrade cerebral perfusion superior to retrograde cerebral perfusion for brain protection during deep hypothermic circulatory arrest? Metabolic evidence from microdialysis.

OBJECTIVES: This study aimed to investigate whether selective antegrade cerebral perfusion or retrograde cerebral perfusion is a better technique for brain protection in deep hypothermic circulatory arrest by obtaining metabolic evidence from microdialysis. DESIGN: Randomized, animal study. SETTING: Assisted circulation laboratory. SUBJECTS: Eighteen piglets of either sex (9.8 ± 3.1 kg). INTERVENTIONS: Animals were randomly assigned to 40 minutes of circulatory arrest at 18°C without cerebral perfusion (deep hypothermic circulatory arrest group, n = 6) or with selective antegrade cerebral perfusion (selective antegrade cerebral perfusion group, n = 6) or retrograde cerebral perfusion (retrograde cerebral perfusion group, n = 6). Reperfusion was continued for 3 hours. MEASUREMENTS AND MAIN RESULTS: Microdialysis (glucose, lactate, pyruvate, and glycerol) variables in the cortex dialysate were measured every 30 minutes. Intracerebral pressure and serum S-100 levels were also monitored. After 3 hours of reperfusion, cortical tissue was harvested for terminal deoxynucleotidyl transferase dUTP nick-end labeling staining. After 40 minutes of circulatory arrest, the deep hypothermic circulatory arrest group presented marked elevations of intracerebral pressure, and serum S-100 levels were higher in the deep hypothermic circulatory arrest group than in the other two groups (p < 0.001, respectively). The selective antegrade cerebral perfusion group exhibited higher glucose, lower lactate, and lower glycerol levels and a lower lactate-to-pyruvate ratio in comparison to the deep hypothermic circulatory arrest group (p < 0.05, respectively); the retrograde cerebral perfusion group had lower lactate and glycerol levels and a lower lactate-to-pyruvate ratio (p < 0.05, respectively) but similar glucose levels compared to deep hypothermic circulatory arrest alone. Furthermore, selective antegrade cerebral perfusion provided better preservation of energy and cell integrity than retrograde cerebral perfusion with higher glucose and lower glycerol levels (p < 0.05, respectively). After 3 hours of reperfusion, fewer apoptotic neurons were found in selective antegrade cerebral perfusion animals than in the other two groups (p < 0.05, respectively). CONCLUSIONS: Both selective antegrade cerebral perfusion and retrograde cerebral perfusion were superior to deep hypothermic circulatory arrest alone during circulatory arrest. Retrograde cerebral perfusion was a moderate technique that had similar advantages with regard to less cerebral edema, better clearance of metabolic waste, and lower levels of biomarkers of injury than selective antegrade cerebral perfusion, but its capacity for energy preservation, maintenance of cellular integrity, and protection against apoptosis was lower than that of selective antegrade cerebral perfusion.

Selective antegrade cerebral perfusion attenuating the TLR4/NF-κB pathway during deep hypothermia circulatory arrest in a pig model.

OBJECTIVES: The alteration of the Toll-like receptor/nuclear factor-kappa B (TLR4/NF-κB) signaling pathway during deep hypothermia circulatory arrest (DHCA) has not yet been defined. The aim of this study was to explore the expression of the TLR4/NF-κB pathway cytokine in cerebral injury resulting from DHCA as well as the effect of selective antegrade cerebral perfusion (SACP) on TLR4/NF-κB pathway expression. METHODS: Twelve pigs were randomly assigned to DHCA alone (n = 6) or DHCA with SACP (n = 6) at 18°C for 80 min. Serum interleukin (IL)-6 was assayed by ELISA. Apoptosis and NF-κB proteins were detected by fluorescence TUNEL and Western blot, respectively. The level of TLR4 mRNA and protein were determined through qRT-PCR and Western blot. RESULTS: The serum IL-6 level of the SACP group was significantly lower than that of the DHCA group at the end of circulation arrest and experimentation. Apoptotic index and NF-κB protein were apparently lower in SACP animals (p < 0.05). Compared to the DHCA group, the levels of TLR4 protein and mRNA in the SACP group were lower with significance (p < 0.05). CONCLUSIONS: The TLR4/NF-κB signaling pathway plays a critical role in the pathogenesis of DHCA cerebral injury. Attenuation of the TLR4/NF-κB inflammatory cytokines probably contributes to the neuroprotective effect of SACP. The TLR4/NF-κB inflammatory signaling pathway may be a novel therapeutic target for developing a new strategy for neuroprotection in DHCA.

Intelligent prediction of Alzheimer's disease via improved multifeature squeeze-and-excitation-dilated residual network.

This study aimed to address the issue of larger prediction errors existing in intelligent predictive tasks related to Alzheimer's disease (AD). A cohort of 487 enrolled participants was categorized into three groups: normal control (138 individuals), mild cognitive impairment (238 patients), and AD (111 patients) in this study. An improved multifeature squeeze-and-excitation-dilated residual network (MFSE-DRN) was proposed for two important AD predictions: clinical scores and conversion probability. The model was characterized as three modules: squeeze-and-excitation-dilated residual block (SE-DRB), multifusion pooling (MF-Pool), and multimodal feature fusion. To assess its performance, the proposed model was compared with two other novel models: ranking convolutional neural network (RCNN) and 3D vision geometrical group network (3D-VGGNet). Our method showed the best performance in the two AD predicted tasks. For the clinical scores prediction, the root-mean-square errors (RMSEs) and mean absolute errors (MAEs) of mini-mental state examination (MMSE) and AD assessment scale-cognitive 11-item (ADAS-11) were 1.97, 1.46 and 4.20, 3.19 within 6 months; 2.48, 1.69 and 4.81, 3.44 within 12 months; 2.67, 1.86 and 5.81, 3.83 within 24 months; 3.02, 2.03 and 5.09, 3.43 within 36 months, respectively. At the AD conversion probability prediction, the prediction accuracies within 12, 24, and 36 months reached to 88.0, 85.5, and 88.4%, respectively. The AD predication would play a great role in clinical applications.

Intelligent quality control of traditional chinese medical tongue diagnosis images based on deep learning.

BACKGROUND: Computer-aided tongue and face diagnosis technology can make Traditional Chinese Medicine (TCM) more standardized, objective and quantified. However, many tongue images collected by the instrument may not meet the standard in clinical applications, which affects the subsequent quantitative analysis. The common tongue diagnosis instrument cannot determine whether the patient has fully extended the tongue or collected the face. OBJECTIVE: This paper proposes an image quality control algorithm based on deep learning to verify the eligibility of TCM tongue diagnosis images. METHODS: We firstly gathered enough images and categorized them into five states. Secondly, we preprocessed the training images. Thirdly, we built a ResNet34 model and trained it by the transfer learning method. Finally, we input the test images into the trained model and automatically filter out unqualified images and point out the reasons. RESULTS: Experimental results show that the model's quality control accuracy rate of the test dataset is as high as 97.06%. Our methods have the strong discriminative power of the learned representation. Compared with previous studies, it can guarantee subsequent tongue image processing. CONCLUSIONS: Our methods can guarantee the subsequent quantitative analysis of tongue shape, tongue state, tongue spirit, and facial complexion.

Advances in systemic immune inflammatory indices in non-small cell lung cancer: A review.

Lung cancer is one of the most prevalent cancers globally, with non-small cell lung cancers constituting the majority. These cancers have a high incidence and mortality rate. In recent years, a growing body of research has demonstrated the intricate link between inflammation and cancer, highlighting that inflammation and cancer are inextricably linked and that inflammation plays a pivotal role in cancer development, progression, and prognosis of cancer. The Systemic Immunoinflammatory Index (SII), comprising neutrophil, lymphocyte, and platelet counts, is a more comprehensive indicator of the host's systemic inflammation and immune status than a single inflammatory index. It is widely used in clinical practice due to its cost-effectiveness, simplicity, noninvasiveness, and ease of acquisition. This paper reviews the impact of SII on the development, progression, and prognosis of non-small cell lung cancer.

Ecotoxicity of hexavalent chromium [Cr (VI)] in soil presents predominate threats to agricultural production with the increase of soil Cr contamination.

The relative severity between chromium (Cr)-mediated ecotoxicity and its bioaccumulation has rarely been compared and evaluated. This study employed pot incubation experiments to simulate the soil environment with increased Cr pollution and study their effects on the growth of crops, including pepper, lettuce, wheat, and rice. Results showed that increasing total Cr presented ascendant ecotoxicity in upland soils when pH > 7.5, and significantly reduced the yield of pepper, lettuce and wheat grain by 0.3-100 %, whereas, this effect was weakened even reversed as the pH decreased. Surprisingly, a series of soils with Cr concentration of 22.7-623.5 mg kg-1 did not cause Cr accumulation in four crops over the Chinese permissible limit. The toxicity of Cr was highly associated with extractable Cr, where Cr (VI) made the greater contributions than Cr (III). Conclusively, the ecotoxicity of Cr poses a greater environmental issue as compared to the bioaccumulation of Cr in crops in upland soils, while extractable Cr (VI) makes the predominant contributions to the ecotoxicity of Cr as the total Cr increased. Our study proposes a synchronous consideration involving total Cr and Cr (VI) as the theoretical basis to establish a more reliable soil quality standard for safe production in China.

Rice straw returning enhances cadmium activation by accelerating iron cycling thus hydroxyl radical production in paddy soils during drainage.

Straw returning is widely found elevating the bioavailability of cadmium (Cd) in paddy soils with unclear biogeochemical mechanisms. Here, a series of microcosm incubation experiments were conducted and spectroscopic and microscopic analyses were employed. The results showed that returning rice straw (RS) efficiently increased amorphous Fe and low crystalline Fe (II) to promote the production of hydroxyl radicals (OH) thus Cd availability in paddy soils during drainage. On the whole, RS increased OH and extractable Cd by 0.2-1.4 and 0.1-3.3 times, respectively. While the addition of RS effectively improved the oxidation rate of structural Fe (II) mineral (i.e., FeS) to enhance soil Cd activation (up to 38.5 %) induced by the increased OH (up to 69.2 %). Additionally, the existence of CO32- significantly increased the efficiency level on OH production and Cd activation, which was attributed to the improved reactivity of Fe (II) by CO32- in paddy soils. Conclusively, this study emphasizes risks of activating soil Cd induced by RS returning-derived OH, providing a new insight into evaluating the safety of straw recycling.

Tissue-derived exosome proteomics identifies promising diagnostic biomarkers for esophageal cancer.

Esophageal cancer (EC) is a fatal digestive disease with a poor prognosis and frequent lymphatic metastases. Nevertheless, reliable biomarkers for EC diagnosis are currently unavailable. Accordingly, we have performed a comparative proteomics analysis on cancer and paracancer tissue-derived exosomes from eight pairs of EC patients using label-free quantification proteomics profiling and have analyzed the differentially expressed proteins through bioinformatics. Furthermore, nano-flow cytometry (NanoFCM) was used to validate the candidate proteins from plasma-derived exosomes in 122 EC patients. Of the 803 differentially expressed proteins discovered in cancer and paracancer tissue-derived exosomes, 686 were up-regulated and 117 were down-regulated. Intercellular adhesion molecule-1 (CD54) was identified as an up-regulated candidate for further investigation, and its high expression in cancer tissues of EC patients was validated using immunohistochemistry, real-time quantitative PCR (RT-qPCR), and western blot analyses. In addition, plasma-derived exosome NanoFCM data from 122 EC patients concurred with our proteomic analysis. The receiver operating characteristic (ROC) analysis demonstrated that the AUC, sensitivity, and specificity values for CD54 were 0.702, 66.13%, and 71.31%, respectively, for EC diagnosis. Small interference (si)RNA was employed to silence the CD54 gene in EC cells. A series of assays, including cell counting kit-8, adhesion, wound healing, and Matrigel invasion, were performed to investigate EC viability, adhesive, migratory, and invasive abilities, respectively. The results showed that CD54 promoted EC proliferation, migration, and invasion. Collectively, tissue-derived exosomal proteomics strongly demonstrates that CD54 is a promising biomarker for EC diagnosis and a key molecule for EC development.

Imaging and pathological characteristics, treatment, and prognosis of pulmonary sequestration-A retrospective study of 13 cases.

OBJECTIVE: This study aimed to summarize and analyze the characteristics of pulmonary sequestration to improve our understanding of this disease. METHODS: Between January 2019 and April 2023, the clinical data of 13 patients with pulmonary sequestration underwent surgical treatment at the First Affiliated Hospital of Gannan Medical University. RESULTS: The male-to-female ratio was 4:9, the age was 0.5 to 60 years, and the average age was 38 ± 19 years. There were 10 and 3 cases of intralobar and extralobar pulmonary sequestration, respectively. Chest enhanced computed tomography (CT) and three-dimensional vascular reconstruction showed that the abnormal blood vessels were derived from the descending thoracic aorta in nine cases and from other blood vessels in four cases. Three patients underwent thoracoscopic lobectomy, two underwent thoracoscopic segmentectomy, and eight underwent thoracoscopic wedge resection. All the patients successfully completed the surgery and were discharged postoperatively. CONCLUSIONS: Some patients with pulmonary sequestration exhibit no obvious symptoms. Patients with clinical symptoms are easily confused for pneumonia, bronchial cysts, lung abscesses, and lung tumors; therefore, patients with pulmonary sequestration are prone to missed diagnosis and misdiagnosis. Currently, enhanced chest CT combined with three-dimensional vascular reconstruction can accurately show the course, branches, and relationship with the mass of the feeding artery. Routine pathological examination is helpful to further clarify the diagnosis of pulmonary sequestration. Minimally invasive thoracoscopic surgery is the preferred treatment for patients with pulmonary sequestration. Surgical resection is safe and feasible, and satisfactory results are typically obtained.

A combined deformable model and medical transformer algorithm for medical image segmentation.

Deep learning-based segmentation models usually require substantial data, and the model usually suffers from poor generalization due to the lack of training data and inefficient network structure. We proposed to combine the deformable model and medical transformer neural network on the image segmentation task to alleviate the aforementioned problems. The proposed method first employs a statistical shape model to generate simulated contours of the target object, and then the thin plate spline is applied to create a realistic texture. Finally, a medical transformer network was constructed to segment three types of medical images, including prostate MR image, heart US image, and tongue color images. The segmentation accuracy of the three tasks achieved 89.97%, 91.90%, and 94.25%, respectively. The experimental results show that the proposed method improves medical image segmentation performance.

Role of exosomal noncoding RNA in esophageal carcinoma.

Esophageal cancer is a common malignant tumor with a high degree of malignancy. Understanding its pathogenesis and identifying early diagnostic biomarkers can significantly improve the prognosis of esophageal cancer patients. Exosomes are small double-membrane vesicles found in various body fluids containing various components (DNA, RNA, and proteins) that mediate intercellular signal communication. Non-coding RNAs are a class of gene transcription products that encode polypeptide functions and are widely detected in exosomes. There is growing evidence that exosomal non-coding RNAs are involved in cancer growth, metastasis and angiogenesis, and can also be used as diagnostic and prognostic markers. This article reviews the recent progress in exosomal non-coding RNAs in esophageal cancer, including research progress, diagnostic value, proliferation, migration, invasion, and drug resistance, provide new ideas for the precise treatment of esophageal cancer.

Role of exosomes in non-small cell lung cancer and EGFR-mutated lung cancer.

As an important mediator of information transfer between cells, exosomes play a unique role in regulating tumor growth, supporting vascular proliferation, tumor invasion, and metastasis. Exosomes are widely present in various body fluids, and therefore they can be used as a potential tool for non-invasive liquid biopsy. The present study reviews the role of exosomes in liquid biopsy, tumor microenvironment formation, and epithelial-mesenchymal transition in non-small cell lung cancer (NSCLC). By targeting epidermal growth factor receptor (EGFR) therapy as a first-line treatment for patients with NSCLC, this study also briefly describes the occurrence of EGRF+ exosomes and the role of exosomes and their contents in non-invasive detection and potential therapeutic targets in EGFR-mutated lung cancer.

Oncogenic zinc finger protein ZNF687 accelerates lung adenocarcinoma cell proliferation and tumor progression by activating the PI3K/AKT signaling pathway.

BACKGROUND: Zinc finger protein 687 (ZNF687) has previously been discovered as a potential oncogene in individuals with giant cell tumors of the bone, acute myeloid leukemia, and hepatocellular carcinoma. However, its role and mechanism in lung adenocarcinoma (LUAD) remain unclear. METHODS: In LUAD cells, tumor, and matched adjacent tissue specimens, quantitative real-time RT- polymerase chain reaction (qRT-PCR), western blotting analyses, and immunohistochemistry staining (IHC) were conducted. Cell counting kit-8 (CCK8) assay, clonogenicity analysis, flow cytometry, and transwell assays were utilized to detect ZNF687 overexpression and knockdown impacts on cell growth, colony formation, cell cycle, migration, and invasion. Bioinformatic studies, qRT-PCR and western blotting studies were employed to validate the underlying mechanisms and signaling pathways implicated in the oncogenic effect of ZNF687. RESULTS: This study demonstrated that ZNF687 expression was elevated in LUAD cells and tissues. Individuals with upregulated ZNF687 had a poorer prognosis than those with downregulatedZNF687 (p < 0.001). ZNF687 overexpression enhanced LUAD growth, migration, invasion and colony formation, and the cell cycle G1-S transition; additionally, it promoted the epithelial-mesenchymal transition (EMT). In contrast, knocking down ZNF687 showed to have the opposite impact. Moreover, these effects were associated with the activity of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) signaling mechanism. CONCLUSION: ZNF687 was upregulated in LUAD, and high ZNF687 expression levels are associated with poor prognoses. The activation of the PI3K/AKT signaling pathway by upregulated ZNF687 increased the proliferation of LUAD cells and tumor progression. ZNF687 may be a beneficial predictive marker and a therapeutic target in LUAD.

Rapid identification of high and low cadmium (Cd) accumulating rice cultivars using machine learning models with molecular markers and soil Cd levels as input data.

Excessive accumulation of cadmium (Cd) in rice grains threatens food safety and human health. Growing low Cd accumulating rice cultivars is an effective approach to produce low-Cd rice. However, field screening of low-Cd rice cultivars is laborious, time-consuming, and subjected to the influence of environment × genotype interactions. In the present study, we investigated whether machine learning-based methods incorporating genotype and soil Cd concentration can identify high and low-Cd accumulating rice cultivars. One hundred and sixty-seven locally adapted high-yielding rice cultivars were grown in three fields with different soil Cd levels and genotyped using four molecular markers related to grain Cd accumulation. We identified sixteen cultivars as stable low-Cd accumulators with grain Cd concentrations below the 0.2 mg kg-1 food safety limit in all three paddy fields. In addition, we developed eight machine learning-based models to predict low- and high-Cd accumulating rice cultivars with genotypes and soil Cd levels as input data. The optimized model classifies low- or high-Cd cultivars (i.e., the grain Cd concentration below or above 0.2 mg kg-1) with an overall accuracy of 76%. These results indicate that machine learning-based classification models constructed with molecular markers and soil Cd levels can quickly and accurately identify the high- and low-Cd accumulating rice cultivars.