Xiqing Li

Background. There is limited knowledge about the role of cancer-associated fibroblasts (CAF) in the tumor microenvironment of triple-negative breast cancer (TNBC). Methods. Three hundred and thirty-five TNBC samples from four datasets were retrieved and analyzed. In order to determine the CAF subtype by combining gene expression profiles, an unsupervised clustering analysis was adopted. The prognosis, enriched pathways, immune cells, immune scores, and tumor purity were compared between CAF subtypes. The genes with the highest importance were selected by bioinformatics analysis. The machine learning model was built to predict the TNBC CAF subtype by these selected genes. Results. TNBC samples were classified into two CAF subtypes (CAF+ and CAF-). The CAF- subtype of TNBC was linked to the longer overall survival and more immune cells than the CAF+ subtype. CAF- and CAF+ were enriched in immune-related pathways and extracellular matrix pathways, respectively. Bioinformatics analysis identified 9 CAF subtype-related markers (ADAMTS12, AEBP1, COL10A1, COL11A1, CXCL11, CXCR6, EDNRA, EPPK1, and WNT7B). We constructed a robust random forest model using these 9 genes, and the area under the curve (AUC) value of the model was 0.921. Conclusion. The current study identified CAF subtypes based on gene expression profiles and found that CAF subtypes have significantly different overall survival, immune cells, and immunotherapy response rates.

Background: MicroRNAs (miRNAs) play key roles in a variety of pathological processes by interacting with their specific target mRNAs for translation repression and may function as oncogenes (oncomiRs) or tumor suppressors (TSmiRs). Therefore, a web server that could predict the regulation relations between miRNAs and small molecules is expected to achieve implications for identifying potential therapeutic targets for anti-tumor drug development.Methods: Upon obtaining positive/known small molecule-miRNA regulation pairs from SM2miR, we generated a multitude of high-quality negative/unknown pairs by leveraging similarities between the small molecule structures. Using the pool of the positive and negative pairs, we created the Dataset1 and Dataset2 datasets specific to up-regulation and down-regulation pairs, respectively. Manifold machine learning algorithms were then employed to construct models of predicting up-regulation and down-regulation pairs on the training portion of pairs in Dataset1 and Dataset2, respectively. Prediction abilities of the resulting models were further examined by discovering potential small molecules to regulate oncogenic miRNAs identified from miRNA sequencing data of endometrial carcinoma samples.Results: The random forest algorithm outperformed four machine-learning algorithms by achieving the highest AUC values of 0.911 for the up-regulation model and 0.896 for the down-regulation model on the testing datasets. Moreover, the down-regulation and up-regulation models yielded the accuracy values of 0.91 and 0.90 on independent validation pairs, respectively. In a case study, our model showed highly-reliable results by confirming all top 10 predicted regulation pairs as experimentally validated pairs. Finally, our predicted binding affinities of oncogenic miRNAs and small molecules bore a close resemblance to the lowest binding energy profiles using molecular docking. Predictions of the final model are freely accessible through the PSRR web server at https://rnadrug.shinyapps.io/PSRR/.Conclusion: Our study provides a novel web server that could effectively predict the regulation of miRNAs expression by small molecules.