Altered expression of Urea Cycle (UC) enzymes occurs in many tumors, resulting a metabolic hallmark termed as UC dysregulation. Polyamines are synthesized from ornithine, and polyamine synthetic genes are elevated in various tumors. However, the underlying deregulations of UC/ polyamine synthesis in cancer remain elusive. Here, we characterized a hypoxia-induced lncRNA LVBU (lncRNA regulation via BCL6/urea cycle) that is highly expressed in colorectal cancer (CRC) and correlates with poor cancer prognosis. Increased LVBU expression promoted CRC cells proliferation, foci formation and tumorigenesis. Further, LVBU regulates urea cycle and polyamine synthesis through BCL6, a negative regulator of p53. Mechanistically, overexpression of LVBU competitively bound miR-10a/miR-34c to protect BCL6 from miR-10a/34c-mediated degradation, which in turn allows BCL6 to block p53-mediated suppression of genes (arginase1 ARG1, ornithine transcarbamylase OTC, ornithine decarboxylase 1 ODC1) involved in UC/polyamine synthesis. Significantly, ODC1 inhibitor attenuated the growth of patient derived xenografts (PDX) that sustain high LVBU levels. Taken together, elevated LVBU can regulate BCL6-p53 signaling axis for systemic UC/polyamine synthesis reprogramming and confers a predilection toward CRC development. Our data demonstrates that further drug development and clinical evaluation of inhibiting UC/polyamine synthesis are warranted for CRC patients with high expression of LVBU.

Facial beauty prediction (FBP) is an important and challenging problem in the fields of computer vision and machine learning. Not only it is easily prone to overfitting due to the lack of large-scale and effective data, but also difficult to quickly build robust and effective facial beauty evaluation models because of the variability of facial appearance and the complexity of human perception. Transfer Learning can be able to reduce the dependence on large amounts of data as well as avoid overfitting problems. Broad learning system (BLS) can be capable of quickly completing models building and training. For this purpose, Transfer Learning was fused with BLS for FBP in this paper. Firstly, a feature extractor is constructed by way of CNNs models based on transfer learning for facial feature extraction, in which EfficientNets are used in this paper, and the fused features of facial beauty extracted are transferred to BLS for FBP, called E-BLS. Secondly, on the basis of E-BLS, a connection layer is designed to connect the feature extractor and BLS, called ER-BLS. Finally, experimental results show that, compared with the previous BLS and CNNs methods existed, the accuracy of FBP was improved by E-BLS and ER-BLS, demonstrating the effectiveness and superiority of the method presented, which can also be widely used in pattern recognition, object detection and image classification.

Numerous studies have demonstrated that individual proteins can moonlight. Eukaryotic Initiation translation factor 3, f subunit (eIF3f) is involved in critical biological functions; however, its role independent of protein translation in regulating colorectal cancer (CRC) is not characterized. Here, it is demonstrated that eIF3f is upregulated in CRC tumor tissues and that both Wnt and EGF signaling pathways are participating in eIF3f's oncogenic impact on targeting phosphoglycerate dehydrogenase (PHGDH) during CRC development. Mechanistically, EGF blocks FBXW7β‐mediated PHGDH ubiquitination through GSK3β deactivation, and eIF3f antagonizes FBXW7β‐mediated PHGDH ubiquitination through its deubiquitinating activity. Additionally, Wnt signals transcriptionally activate the expression of eIF3f, which also exerts its deubiquitinating activity toward MYC, thereby increasing MYC‐mediated PHGDH transcription. Thereby, both impacts allow eIF3f to elevate the expression of PHGDH, enhancing Serine–Glycine–One–Carbon (SGOC) signaling pathway to facilitate CRC development. In summary, the study uncovers the intrinsic role and underlying molecular mechanism of eIF3f in SGOC signaling, providing novel insight into the strategies to target eIF3f‐PHGDH axis in CRC.

Facial Beauty Prediction (FBP) is an important and challenging problem in the field of computer vision and machine learning. Not only it is easily prone to over-fitting due to the lack of large-scale and effective data, but also difficult to quickly build robust and effective face beauty evaluation models because of the variability of facial appearance and the complexity of human perception. Transfer learning can be able to reduce the dependence on large amounts of data as well as avoid overfitting problems. Broad Learning System (BLS) can be capable of quickly completing models building and training. For this purpose, transfer learning was fused with BLS for facial beauty prediction in this paper. Firstly, a feature extractor is constructed by way of CNN model based on transfer learning for facial feature extraction, in which EfficientNet is used in this paper, and the facial features extracted are transferred to BLS for facial beauty prediction, called E-BLS. Then, on the basis of E-BLS, a connection layer is designed to connect the feature extractor and BLS, called ER-BLS. Experimental results show that, compared with the previous BLS and CNN methods existed, the accuracy of facial beauty prediction was improved by E-BLS and ER-BLS, indicating the effectiveness of the method presented, which can also be widely used in pattern recognition, object detection and image classification, etc.