Ultrasound deep learning radiomics and clinical machine learning models to predict low nuclear grade, ER, PR, and HER2 receptor status in pure ductal carcinoma in situ.

Background: Low nuclear grade ductal carcinoma in situ (DCIS) patients can adopt proactive management strategies to avoid unnecessary surgical resection. Different personalized treatment modalities may be selected based on the expression status of molecular markers, which is also predictive of different outcomes and risks of recurrence. DCIS ultrasound findings are mostly non mass lesions, making it difficult to determine boundaries. Currently, studies have shown that models based on deep learning radiomics (DLR) have advantages in automatic recognition of tumor contours. Machine learning models based on clinical imaging features can explain the importance of imaging features. Methods: The available ultrasound data of 349 patients with pure DCIS confirmed by surgical pathology [54 low nuclear grade, 175 positive estrogen receptor (ER+), 163 positive progesterone receptor (PR+), and 81 positive human epidermal growth factor receptor 2 (HER2+)] were collected. Radiologists extracted ultrasonographic features of DCIS lesions based on the 5th Edition of Breast Imaging Reporting and Data System (BI-RADS). Patient age and BI-RADS characteristics were used to construct clinical machine learning (CML) models. The RadImageNet pretrained network was used for extracting radiomics features and as an input for DLR modeling. For training and validation datasets, 80% and 20% of the data, respectively, were used. Logistic regression (LR), support vector machine (SVM), random forest (RF), and eXtreme Gradient Boosting (XGBoost) algorithms were performed and compared for the final classification modeling. Each task used the area under the receiver operating characteristic curve (AUC) to evaluate the effectiveness of DLR and CML models. Results: In the training dataset, low nuclear grade, ER+, PR+, and HER2+ DCIS lesions accounted for 19.20%, 65.12%, 61.21%, and 30.19%, respectively; the validation set, they consisted of 19.30%, 62.50%, 57.14%, and 30.91%, respectively. In the DLR models we developed, the best AUC values for identifying features were 0.633 for identifying low nuclear grade, completed by the XGBoost Classifier of ResNet50; 0.618 for identifying ER, completed by the RF Classifier of InceptionV3; 0.755 for identifying PR, completed by the XGBoost Classifier of InceptionV3; and 0.713 for identifying HER2, completed by the LR Classifier of ResNet50. The CML models had better performance than DLR in predicting low nuclear grade, ER+, PR+, and HER2+ DCIS lesions. The best AUC values by classification were as follows: for low nuclear grade by RF classification, AUC: 0.719; for ER+ by XGBoost classification, AUC: 0.761; for PR+ by XGBoost classification, AUC: 0.780; and for HER2+ by RF classification, AUC: 0.723. Conclusions: Based on small-scale datasets, our study showed that the DLR models developed using RadImageNet pretrained network and CML models may help predict low nuclear grade, ER+, PR+, and HER2+ DCIS lesions so that patients benefit from hierarchical and personalized treatment.

Precise prediction of phase-separation key residues by machine learning.

Understanding intracellular phase separation is crucial for deciphering transcriptional control, cell fate transitions, and disease mechanisms. However, the key residues, which impact phase separation the most for protein phase separation function have remained elusive. We develop PSPHunter, which can precisely predict these key residues based on machine learning scheme. In vivo and in vitro validations demonstrate that truncating just 6 key residues in GATA3 disrupts phase separation, enhancing tumor cell migration and inhibiting growth. Glycine and its motifs are enriched in spacer and key residues, as revealed by our comprehensive analysis. PSPHunter identifies nearly 80% of disease-associated phase-separating proteins, with frequent mutated pathological residues like glycine and proline often residing in these key residues. PSPHunter thus emerges as a crucial tool to uncover key residues, facilitating insights into phase separation mechanisms governing transcriptional control, cell fate transitions, and disease development.

DASES: a database of alternative splicing for esophageal squamous cell carcinoma.

Esophageal carcinoma ranks as the sixth leading cause of cancer-related mortality globally, with esophageal squamous cell carcinoma (ESCC) being particularly prevalent among Asian populations. Alternative splicing (AS) plays a pivotal role in ESCC development and progression by generating diverse transcript isoforms. However, the current landscape lacks a specialized database focusing on alternative splicing events (ASEs) derived from a large number of ESCC cases. Additionally, most existing AS databases overlook the contribution of long non-coding RNAs (lncRNAs) in ESCC molecular mechanisms, predominantly focusing on mRNA-based ASE identification. To address these limitations, we deployed DASES (http://www.hxdsjzx.cn/DASES). Employing a combination of publicly available and in-house ESCC RNA-seq datasets, our extensive analysis of 346 samples, with 93% being paired tumor and adjacent non-tumor tissues, led to the identification of 257 novel lncRNAs in esophageal squamous cell carcinoma. Leveraging a paired comparison of tumor and adjacent normal tissues, DASES identified 59,094 ASEs that may be associated with ESCC. DASES fills a critical gap by providing comprehensive insights into ASEs in ESCC, encompassing lncRNAs and mRNA, thus facilitating a deeper understanding of ESCC molecular mechanisms and serving as a valuable resource for ESCC research communities.

The Causal Association of Irritable Bowel Syndrome with Multiple Disease Outcomes: A Phenome-Wide Mendelian Randomization Study.

BACKGROUND: This study aimed to identify novel associations between irritable bowel syndrome (IBS) and a broad range of outcomes. METHODS: In total, 346,352 white participants in the U.K. Biobank were randomly divided into two halves, in which a genome-wide association study (GWAS) of IBS and a polygenic risk score (PRS) analysis of IBS using GWAS summary statistics were conducted, respectively. A phenome-wide association study (PheWAS) based on the PRS of IBS was performed to identify disease outcomes associated with IBS. Then, the causalities of these associations were tested by both one-sample (individual-level data in U.K. Biobank) and two-sample (publicly available summary statistics) Mendelian randomization (MR). Sex-stratified PheWAS-MR analyses were performed in male and female, separately. RESULTS: Our PheWAS identified five diseases associated with genetically predicted IBS. Conventional MR confirmed these causal associations between IBS and depression (OR: 1.07, 95%CI: 1.01-1.14, p = 0.02), diverticular diseases of the intestine (OR: 1.13, 95%CI: 1.08-1.19, p = 3.00 × 10-6), gastro-esophageal reflux disease (OR: 1.09, 95%CI: 1.05-1.13, p = 3.72 × 10-5), dyspepsia (OR: 1.21, 95%CI: 1.13-1.30, p = 9.28 × 10-8), and diaphragmatic hernia (OR: 1.10, 95%CI: 1.05-1.15, p = 2.75 × 10-5). The causality of these associations was observed in female only, but not men. CONCLUSIONS: Increased risks of IBS is found to cause a series of disease outcomes. Our findings support further investigation on the clinical relevance of increased IBS risks with mental and digestive disorders.

The impact of age-specific childhood body-mass index on adult cardiometabolic traits: a Mendelian randomization study.

Objective: To evaluate the causal relationship between childhood body-mass index (BMI) at different ages and adult cardiometabolic traits. Methods: We retrieved genetic instrument variables (IVs) for exposures (standardized BMI at newborn, infant, toddler and late childhood), cardiometabolic traits and potential confounders or mediators (adult BMI, SHBG, testosterone and age at menarche) from the corresponding genome-wide association analysis. We performed univariate and multivariable Mendelian randomization (MR) to dissect associations between age-specific childhood BMI and adult cardiometabolic outcomes. Odds ratio was used to present the direction of the causal association. Results: In univariate MR, higher newborn BMI was causally associated with reduced risk for type 2 diabetes in women. Late childhood BMI was associated with increased risk for female diabetes and coronary artery disease (CAD), myocardial infarction (MI), and chronic kidney disease (CKD) in general population. Among these associations, only association between late childhood BMI with MI remained significant after adjusting for adult male BMI and sex hormones, (OR = 1.120, 95% CI 1.023-1.226, p = 0.014). Besides, in multivariable MR, we found evidence for causal association between newborn BMI with reduced risk for CAD (OR = 0.862, 95% CI 0.751-0.989, p = 0.034) and MI (OR = 0.864, 95% CI 0.752-0.991, p = 0.037) in men. No obvious impact of infant or toddler BMI was identified on the above-mentioned diseases. For continuous cardiometabolic traits, in all age epochs except infant, higher BMI was associated with increased level of fasting glucose in women. Conclusion: BMI at birth and late childhood exerts different impact on adult cardiometabolic diseases, while BMI at infant and toddler ages is not causally associated with these outcomes. The effect of childhood BMI may be influenced by sex disparities.