Pan-cancer integrated bioinformatic analysis of RNA polymerase subunits reveal RNA Pol I member CD3EAP regulates cell growth by modulating autophagy.

Transcription is a crucial stage in gene expression. An integrated study of 34 RNA polymerase subunits (RNAPS) in the six most frequent cancer types identified several genetic and epigenetic modification. We discovered nine mutant RNAPS with a mutation frequency of more than 1% in at least one tumor type. POLR2K and POLR2H were found to be amplified and overexpressed, whereas POLR3D was deleted and downregulated. Multiple RNAPS were also observed to be regulated by variations in promoter methylation. 5-Aza-2-deoxycytidine mediated re-expression in cell lines verified methylation-driven inhibition of POLR2F and POLR2L expression in BRCA and NSCLC, respectively. Next, we showed that CD3EAP, a Pol I subunit, was overexpressed in all cancer types and was associated with worst survival in breast, liver, lung, and prostate cancers. The knockdown studies showed that CD3EAP is required for cell proliferation and induces autophagy but not apoptosis. Furthermore, autophagy inhibition rescued the cell proliferation in CD3EAP knockdown cells. CD3EAP expression correlated with S and G2 phase cell cycle regulators, and CD3EAP knockdown inhibited the expression of S and G2 CDK/cyclins. We also identified POLR2D, an RNA pol II subunit, as a commonly overexpressed and prognostic gene in multiple cancers. POLR2D knockdown also decreased cell proliferation. POLR2D is related to the transcription of just a subset of RNA POL II transcribe genes, indicating a distinct role. Taken together, we have shown the genetic and epigenetic regulation of RNAPS genes in most common tumors. We have also demonstrated the cancer-specific function of CD3EAP and POLR2D genes.

Integrative gene expression analysis for the diagnosis of Parkinson's disease using machine learning and explainable AI.

Parkinson's disease (PD) is a progressive neurodegenerative disorder. Various symptoms and diagnostic tests are used in combination for the diagnosis of PD; however, accurate diagnosis at early stages is difficult. Blood-based markers can support physicians in the early diagnosis and treatment of PD. In this study, we used Machine Learning (ML) based methods for the diagnosis of PD by integrating gene expression data from different sources and applying explainable artificial intelligence (XAI) techniques to find the significant set of gene features contributing to diagnosis. We utilized the Least Absolute Shrinkage and Selection Operator (LASSO), and Ridge regression for the feature selection process. We utilized state-of-the-art ML techniques for the classification of PD cases and healthy controls. Logistic regression and Support Vector Machine showed the highest diagnostic accuracy. SHapley Additive exPlanations (SHAP) based global interpretable model-agnostic XAI method was utilized for the interpretation of the Support Vector Machine model. A set of significant biomarkers that contributed to the diagnosis of PD were identified. Some of these genes are associated with other neurodegenerative diseases. Our results suggest that the utilization of XAI can be useful in making early therapeutic decisions for the treatment of PD. The integration of datasets from different sources made this model robust. We believe that this research article will be of interest to clinicians as well as computational biologists in translational research.

A comprehensive survey on computational learning methods for analysis of gene expression data.

Computational analysis methods including machine learning have a significant impact in the fields of genomics and medicine. High-throughput gene expression analysis methods such as microarray technology and RNA sequencing produce enormous amounts of data. Traditionally, statistical methods are used for comparative analysis of gene expression data. However, more complex analysis for classification of sample observations, or discovery of feature genes requires sophisticated computational approaches. In this review, we compile various statistical and computational tools used in analysis of expression microarray data. Even though the methods are discussed in the context of expression microarrays, they can also be applied for the analysis of RNA sequencing and quantitative proteomics datasets. We discuss the types of missing values, and the methods and approaches usually employed in their imputation. We also discuss methods of data normalization, feature selection, and feature extraction. Lastly, methods of classification and class discovery along with their evaluation parameters are described in detail. We believe that this detailed review will help the users to select appropriate methods for preprocessing and analysis of their data based on the expected outcome.

An epithelial-mesenchymal plasticity signature identifies two novel LncRNAs with the opposite regulation.

The epithelial to mesenchymal transition (EMT) is crucial for cancer progression and chemoresistance. EMT is a dynamic process with multiple phases that change cell migration and invasion activity. We used pan-cancer expression data to find 14-LncRNAs that had a high correlation with the EMT markers VIM, CDH1, FN1, SNAI1, and SNAI2. The expression of 14 EMT-associated LncRNA, which also showed high cancer specificity, was used to calculate the pan-cancer EMT score. The EMT score was then applied to the 32 cancer types to classify them as epithelial, epithelial-mesenchymal, mesenchymal-epithelial, or mesenchymal tumors. We discovered that the EMT score is a poor prognostic predictor and that as tumor mesenchymal nature increased, patient survival decreased. We also showed that the cell of origin did not influence the EMT nature of tumors. Pathway analysis employing protein expression data revealed that the PI3K pathway is the most crucial in determining the EMTness of tumors. Further, we divided CCLE-cell lines into EMT classes and discovered that mesenchymal cells, which exhibited higher PI3K pathway activation, were more sensitive to PI3K inhibitors than epithelial cells. We identified Linc01615 as a mesenchymal LncRNA whose expression significantly correlated with survival in several cancer types. We showed that Linc01615 is regulated by the TGFß-STAT3 pathway in a feedback loop. Knockdown of Linc01615 inhibited cell proliferation and migration by regulating the PI3K pathway and mesenchymal markers. We also identified RP4-568C11.4 as an epithelial cancer marker. We showed that knocking down RP4-568C11.4 decreased cell growth but not migration. In addition, we discovered that ESR1 regulates RP4-5681C11.4 in breast cancer. Taken together, we have developed a pan-cancer EMT signature. Also, we found two new LncRNAs that have different effects on cancer development and EMT.

Comparison of machine learning and deep learning techniques in promoter prediction across diverse species.

Gene promoters are the key DNA regulatory elements positioned around the transcription start sites and are responsible for regulating gene transcription process. Various alignment-based, signal-based and content-based approaches are reported for the prediction of promoters. However, since all promoter sequences do not show explicit features, the prediction performance of these techniques is poor. Therefore, many machine learning and deep learning models have been proposed for promoter prediction. In this work, we studied methods for vector encoding and promoter classification using genome sequences of three distinct higher eukaryotes viz. yeast (Saccharomyces cerevisiae), A. thaliana (plant) and human (Homo sapiens). We compared one-hot vector encoding method with frequency-based tokenization (FBT) for data pre-processing on 1-D Convolutional Neural Network (CNN) model. We found that FBT gives a shorter input dimension reducing the training time without affecting the sensitivity and specificity of classification. We employed the deep learning techniques, mainly CNN and recurrent neural network with Long Short Term Memory (LSTM) and random forest (RF) classifier for promoter classification at k-mer sizes of 2, 4 and 8. We found CNN to be superior in classification of promoters from non-promoter sequences (binary classification) as well as species-specific classification of promoter sequences (multiclass classification). In summary, the contribution of this work lies in the use of synthetic shuffled negative dataset and frequency-based tokenization for pre-processing. This study provides a comprehensive and generic framework for classification tasks in genomic applications and can be extended to various classification problems.

Sunlight induced biosynthesis of silver nanoparticle from the bark extract of Amentotaxus assamica D.K. Ferguson and its antibacterial activity against Escherichia coli and Staphylococcus aureus.

Green synthesis of nanoparticles has gained importance due to its eco-friendly, low toxicity and cost effective nature. This study deals with the biosynthesis of silver nanoparticles (AgNPs) from the bark extract of Amentotaxus assamica. The AgNPs have been synthesised by reducing the silver ions into stable AgNPs using the bark extract of Amentotaxus assamica under the influence of sunlight irradiation. The characterisation of the biosynthesised AgNPs was carried out by UV-vis spectroscopy, X-ray diffraction analysis (XRD), Fourier transform infrared spectroscopy, scanning electron microscopy (SEM) and energy dispersive X-ray analysis. The UV-vis spectrum showed a broad peak at 472 nm. Also, the XRD confirmed the crystalline structure of the AgNPs. Moreover, the SEM analysis revealed that the biosynthesised AgNPs were spherical in shape. Also, dynamic light scattering techniques were used to evaluate the size distribution profile of the biosynthesised AgNPs. Furthermore, the biosynthesised AgNPs showed a prominent inhibitory effect against both Escherichia coli (MTCC 111) and Staphylococcus aureus (MTCC 97). Thus the biosynthesis of AgNPs from the bark extract of Amentotaxus assamica is found to eco-friendly way of producing AgNPs compared to chemical method.