Cautions about the reliability of pairwise gene correlations based on expression data

miRbiom: Machine-learning on Bayesian causal nets of RBP-miRNA interactions successfully predicts miRNA profiles

Formation of mature miRNAs and their expression is a highly controlled process. It is very much dependent upon the post-transcriptional regulatory events. Recent findings suggest that several RNA binding proteins beyond Drosha/Dicer are involved in the processing of miRNAs. Deciphering of conditional networks for these RBP-miRNA interactions may help to reason the spatio-temporal nature of miRNAs which can also be used to predict miRNA profiles. In this direction, >25TB of data from different platforms were studied (CLIP-seq/RNA-seq/miRNA-seq) to develop Bayesian causal networks capable of reasoning miRNA biogenesis. The networks ably explained the miRNA formation when tested across a large number of conditions and experimentally validated data. The networks were modeled into an XGBoost machine learning system where expression information of the network components was found capable to quantitatively explain the miRNAs formation levels and their profiles. The models were developed for 1,204 human miRNAs whose accurate expression level could be detected directly from the RNA-seq data alone without any need of doing separate miRNA profiling experiments like miRNA-seq or arrays. A first of its kind, miRbiom performed consistently well with high average accuracy (91%) when tested across a large number of experimentally established data from several conditions. It has been implemented as an interactive open access web-server where besides finding the profiles of miRNAs, their downstream functional analysis can also be done. miRbiom will help to get an accurate prediction of human miRNAs profiles in the absence of profiling experiments and will be an asset for regulatory research areas. The study also shows the importance of having RBP interaction information in better understanding the miRNAs and their functional projectiles where it also lays the foundation of such studies and software in future.

Silver nanoparticles-induced H2O2 triggers apoptosis-like death and is associated with dinF in Escherichia coli

Silver nanoparticles (AgNPs) are the most widely used nanomaterials as antimicrobial agents. AgNPs have been shown to inhibit the growth of and induce apoptosis-like death in Escherichia coli. However, the precise mechanism of AgNPs-induced apoptosis-like death and association with DNA damage-inducible protein F (dinF), a gene of SOS response, is unknown. Here, AgNPs-contributing depletion of intracellular glutathione levels and deactivation of glutathione peroxidase were shown. This step, indicating disruption of the antioxidant system, resulted in overall oxidative stress. Furthermore, DNA oxidation was accompanied, leading to DNA fragmentation. In addition, AgNPs appeared to induce apoptosis-like death via the SOS response. We used sodium pyruvate - an H₂O₂ quencher - to study the contribution of H₂O₂, which showed attenuation of AgNPs-induced DNA damage, SOS response, and apoptosis-like death. In dinF mutant, the strain showed a higher degree of DNA damage and apoptotic features. In conclusion, AgNPs mediate apoptosis-like cell death by H₂O₂-induced oxidative DNA damage. Furthermore, our result demonstrates that dinF participates in this process, which further supports that AgNPs induces SOS response. Our findings may contribute to expanding the new applications of AgNP-based nanomaterials in biomedical fields.

IMPROVED PERFORMANCE OF GENE SET ANALYSIS ON GENOME-WIDE TRANSCRIPTOMICS DATA WHEN USING GENE ACTIVITY STATE ESTIMATES

Gene set analysis methods continue to be a popular and powerful method of evaluating genome-wide transcriptomics data. These approach require a priori grouping of genes into biologically meaningful sets, and then conducting downstream analyses at the set (instead of gene) level of analysis. Gene set analysis methods have been shown to yield more powerful statistical conclusions than single-gene analyses due to both reduced multiple testing penalties and potentially larger observed effects due to the aggregation of effects across multiple genes in the set. Traditionally, gene set analysis methods have been applied directly to normalized, log-transformed, transcriptomics data. Recently, efforts have been made to transform transcriptomics data to scales yielding more biologically interpretable results. For example, recently proposed models transform log-transformed transcriptomics data to a confidence metric (ranging between 0 and 100%) that a gene is active (roughly speaking, that the gene product is part of an active cellular mechanism). In this manuscript, we demonstrate, on both real and simulated transcriptomics data, that tests for differential expression between sets of genes using are typically more powerful when using gene activity state estimates as opposed to log-transformed gene expression data. Our analysis suggests further exploration of techniques to transform transcriptomics data to meaningful quantities for improved downstream inference.

A Bayesian Framework for the Classification of Microbial Gene Activity States

Numerous methods for classifying gene activity states based on gene expression data have been proposed for use in downstream applications, such as incorporating transcriptomics data into metabolic models in order to improve resulting flux predictions. These methods often attempt to classify gene activity for each gene in each experimental condition as belonging to one of two states: active (the gene product is part of an active cellular mechanism) or inactive (the cellular mechanism is not active). These existing methods of classifying gene activity states suffer from multiple limitations, including enforcing unrealistic constraints on the overall proportions of active and inactive genes, failing to leverage a priori knowledge of gene co-regulation, failing to account for differences between genes, and failing to provide statistically meaningful confidence estimates. We propose a flexible Bayesian approach to classifying gene activity states based on a Gaussian mixture model. The model integrates genome-wide transcriptomics data from multiple conditions and information about gene co-regulation to provide activity state confidence estimates for each gene in each condition. We compare the performance of our novel method to existing methods on both simulated data and real data from 907 E. coli gene expression arrays, as well as a comparison with experimentally measured flux values in 29 conditions, demonstrating that our method provides more consistent and accurate results than existing methods across a variety of metrics.