The increasing amount of transcriptomic data has brought to light vast numbers of potential novel RNA transcripts. Accurately distinguishing novel long non-coding RNAs (lncRNAs) from protein-coding messenger RNAs (mRNAs) has challenged bioinformatic tool developers. Most recently, tools implementing deep learning architectures have been developed for this task, with the potential of discovering sequence features and their interactions still not surfaced in current knowledge. We compared the performance of deep learning tools with other predictive tools that are currently used in lncRNA coding potential prediction. A total of 15 tools representing the variety of available methods were investigated. In addition to known annotated transcripts, we also evaluated the use of the tools in actual studies with real-life data. The robustness and scalability of the tools' performance was tested with varying sized test sets and test sets with different proportions of lncRNAs and mRNAs. In addition, the ease-of-use for each tested tool was scored. Deep learning tools were top performers in most metrics and labelled transcripts similarly with each other in the real-life dataset. However, the proportion of lncRNAs and mRNAs in the test sets affected the performance of all tools. Computational resources were utilized differently between the top-ranking tools, thus the nature of the study may affect the decision of choosing one well-performing tool over another. Nonetheless, the results suggest favouring the novel deep learning tools over other tools currently in broad use.

Large quantity of variants of uncertain significance (VUS) has been identified in cancer predisposition genes, but classification of VUS remains a big challenge. We proposed that the impact of VUS on protein structure stability can be used to identify these with deleterious effects by using molecular dynamics simulation (MDS)-based approach and developed a MDS-based method for missense VUS classification. In the current study, we applied the system to classify the missense VUS in BRCA2. BRCA2 plays an important role in maintaining genome stability by repairing double-strand DNA damage through homologous recombination. BRCA2 BRC repeats bring RAD51 from cytoplasm to the break sites in nucleus to initiate the repairing process. Missense variants in BRCA2 BRC repeats can interfere the interaction between BRCA2 and RAD51, impair double-strand break repair, cause genome instability and increase cancer risk. We characterized the missense VUS in BRCA2 BRC4 repeat, the primary site of BRCA2 interacting with RAD51. Based on the well-determined BRC4 structure, we applied MDS to measure the impact of BRC4 missense VUS on the stability of BRC4 structure by testing the equilibrium state, flexibility, compactness, hydrogen bonds and surface accessibility. Of the 46 missense VUS analyzed, we were able to differentiate them into 24 Deleterious and 22 Tolerated variants. Comparison between the MDS-based and other 24 existing computational methods for variant classification showed that the MDS-based approach is highly sensitive and specific for classifying missense VUS in cancer predisposition genes.

Triple-negative breast cancer (TNBC) is the breast cancer subtype with the highest fatality rate, and it seriously threatens women's health. Recent studies found that the level of immune cell infiltration in TNBC was associated with tumor progression and prognosis. However, due to practical constraints, most of these TNBC immune infiltration studies only used a small number of patient samples and a few immune cell types. Therefore, it is necessary to integrate more TNBC patient samples and immune cell types to comprehensively study immune infiltration in TNBC to contribute to the prognosis and treatment of patients. In this study, 12 TNBC datasets were integrated and an extensive collection of 182 gene sets with immune-related signatures were included to comprehensively investigate tumor immune microenvironment of TNBC. A single sample gene set enrichment analysis was performed to calculate the infiltration score of each immune-related signature in each patient, and an immune-related risk scoring model for TNBC was constructed to accurately assess patient prognosis. Significant differences were found in immunogenomic landscape between different immune risk subtypes. In addition, the immunotherapy response and chemotherapy drug sensitivity of patients with different immune risk subtypes were also analyzed. The results showed that there were significant differences in these characteristics. Finally, a prediction model for immune risk subtypes of TNBC patients was constructed to accurately predict patients with unknown subtypes. Based on the aforementioned findings, we believed that the immune-related risk score constructed in this study can assist in providing personalized medicine to TNBC patients.