OperonSEQer: A set of machine-learning algorithms with threshold voting for detection of operon pairs using short-read RNA-sequencing data

Operon prediction in prokaryotes is critical not only for understanding the regulation of endogenous gene expression, but also for exogenous targeting of genes using newly developed tools such as CRISPR-based gene modulation. A number of methods have used transcriptomics data to predict operons, based on the premise that contiguous genes in an operon will be expressed at similar levels. While promising results have been observed using these methods, most of them do not address uncertainty caused by technical variability between experiments, which is especially relevant when the amount of data available is small. In addition, many existing methods do not provide the flexibility to determine the stringency with which genes should be evaluated for being in an operon pair. We present OperonSEQer, a set of machine learning algorithms that uses the statistic and p-value from a non-parametric analysis of variance test (Kruskal-Wallis) to determine the likelihood that two adjacent genes are expressed from the same RNA molecule. We implement a voting system to allow users to choose the stringency of operon calls depending on whether your priority is high recall or high specificity. In addition, we provide the code so that users can retrain the algorithm and re-establish hyperparameters based on any data they choose, allowing for this method to be expanded as additional data is generated. We show that our approach detects operon pairs that are missed by current methods by comparing our predictions to publicly available long-read sequencing data. OperonSEQer therefore improves on existing methods in terms of accuracy, flexibility, and adaptability.

Transcriptomics of receptive endometrium in women with sonographic features of adenomyosis

Background Women with uterine adenomyosis seeking assisted reproduction have been associated with compromised endometrial receptivity to embryo implantation. To understand the mechanisms involved in this process, we aimed to compare endometrial transcriptome profiles during the window of implantation (WOI) between women with and without adenomyosis. Methods We obtained endometrial biopsies LH-timed to the WOI from women with sonographic features of adenomyosis (n=10) and controls (n=10). Isolated RNA samples were subjected to RNA sequencing (RNA-seq) by the Illumina NovaSeq 6000 platform and endometrial receptivity classification with a molecular tool for menstrual cycle phase dating (beREADY®, CCHT). The program language R and Bioconductor packages were applied to analyse RNA-seq data in the setting of the result of accurate endometrial dating. To suggest robust candidate pathways, the identified differentially expressed genes (DEGs) associated with the adenomyosis group in the receptive phase were further integrated with 151, 173 and 42 extracted genes from published studies that were related to endometrial receptivity in healthy uterus, endometriosis and adenomyosis, respectively. Enrichment analyses were performed using Cytoscape ClueGO and CluePedia apps. Results Out of 20 endometrial samples, 2 were dated to the early receptive phase, 13 to the receptive phase and 5 to the late receptive phase. Comparison of the transcriptomics data from all 20 samples provided 909 DEGs (p<0.05; nonsignificant after adjusted p value) in the adenomyosis group but only 4 enriched pathways (Bonferroni p value < 0.05). The analysis of 13 samples only dated to the receptive phase provided suggestive 382 DEGs (p<0.05; nonsignificant after adjusted p value) in the adenomyosis group, leading to 33 enriched pathways (Bonferroni p value < 0.05). These included pathways were already associated with endometrial biology, such as “Expression of interferon (IFN)-induced genes” and “Response to IFN-alpha”. Data integration revealed pathways indicating a unique effect of adenomyosis on endometrial molecular organization (e.g., “Expression of IFN-induced genes”) and its interference with endometrial receptivity establishment (e.g., “Extracellular matrix organization” and “Tumour necrosis factor production”). Conclusions Accurate endometrial dating and RNA-seq analysis resulted in the identification of altered response to IFN signalling as the most promising candidate of impaired uterine receptivity in adenomyosis.

The Mammalian Kalrn Locus Gives Rise to Several Novel Long Non-coding RNAs

The Kalrn gene encodes several multi-domain protein isoforms that localise to neuronal synapses, and play dynamic roles in shaping axonal outgrowth, dendrite morphology and dendritic spine re-modelling. The genomic locus is implicated in several neurodevelopmental and neuropsychiatric diseases including autism, schizophrenia and bipolar disease. Mutations in the coding regions, inherited in a classical Mendelian manner, have also been implicated in certain forms of autism and intellectual disability. At the molecular level, the protein isoforms, encoded by reported transcript isoforms, share some core domains arising from the central exons, while other domains, especially towards the C terminal may be selectively incorporated. This heterogeneity seems to confer the ability to grow and retract dendritic spines, thus making Kalirin a critical and dynamic player in dendritogenesis. We have previously shown that in the zebrafish genome, a novel brain specific non-coding RNA arising from the 5’ end of the Kalirin gene, durga regulates neuronal morphology. In search of the mammalian equivalent, we characterized the mammalian Kalrn loci in detail, annotating multiple novel non-coding RNAs, including linear and circular variants, through analysis of transcriptomics data and experimental approaches. By comparing the mouse and human loci and studying the expression of the novel lncRNAs arising from the locus during differentiation of primary cortical neurons in culture, we show that certain non-coding RNAs arising from the locus show a temporal expression profile that coincides with a subset of Kalirin protein coding isoforms. In humans, mouse and zebrafish the 5’end of the Kalrn locus gives rise to a chromatin associated lncRNA that is present in adult ovaries besides being expressed during brain development and in certain regions of the adult brain. Besides correcting some of the annotations available in public databases, we propose that this lncRNA arising from the 5’end of the Kalrn locus is the mammalian ortholog of zebrafish lncRNA durga.

Joint dimension reduction and clustering analysis for single-cell RNA-seq and spatial transcriptomics data

Dimension reduction and (spatial) clustering are two key steps for the analysis of both single-cell RNA-sequencing (scRNA-seq) and spatial transcriptomics data collected from different platforms. Most existing methods perform dimension reduction and (spatial) clustering sequentially, treating them as two consecutive stages in tandem analysis. However, the low-dimensional embeddings estimated in the dimension reduction step may not necessarily be relevant to the class labels inferred in the clustering step and thus may impair the performance of the clustering and other downstream analysis. Here, we develop a computation method, DR-SC, to perform both dimension reduction and (spatial) clustering jointly in a unified framework. Joint analysis in DR-SC ensures accurate (spatial) clustering results and effective extraction of biologically informative low-dimensional features. Importantly, DR-SC is not only applicable for cell type clustering in scRNA-seq studies but also applicable for spatial clustering in spatial transcriptimics that characterizes the spatial organization of the tissue by segregating it into multiple tissue structures. For spatial transcriptoimcs analysis, DR-SC relies on an underlying latent hidden Markov random field model to encourage the spatial smoothness of the detected spatial cluster boundaries. We also develop an efficient expectation-maximization algorithm based on an iterative conditional mode. DR-SC is not only scalable to large sample sizes, but is also capable of optimizing the spatial smoothness parameter in a data-driven manner. Comprehensive simulations show that DR-SC outperforms existing clustering methods such as Seurat and spatial clustering methods such as BayesSpace and SpaGCN and extracts more biologically relevant features compared to the conventional dimension reduction methods such as PCA and scVI. Using 16 benchmark scRNA-seq datasets, we demonstrate that the low-dimensional embeddings and class labels estimated from DR-SC lead to improved trajectory inference. In addition, analyzing three published scRNA-seq and spatial transcriptomics data in three platforms, we show DR-SC can improve both the spatial and non-spatial clustering performance, resolving a low-dimensional representation with improved visualization, and facilitate the downstream analysis such as trajectory inference.

HiPSC-Derived Hepatocyte-like Cells Can Be Used as a Model for Transcriptomics-Based Study of Chemical Toxicity

Traditional toxicity risk assessment approaches have until recently focussed mainly on histochemical readouts for cell death. Modern toxicology methods attempt to deduce a mechanistic understanding of pathways involved in the development of toxicity, by using transcriptomics and other big data-driven methods such as high-content screening. Here, we used a recently described optimised method to differentiate human induced pluripotent stem cells (hiPSCs) to hepatocyte-like cells (HLCs), to assess their potential to classify hepatotoxic and non-hepatotoxic chemicals and their use in mechanistic toxicity studies. The iPSC-HLCs could accurately classify chemicals causing acute hepatocellular injury, and the transcriptomics data on treated HLCs obtained by TempO-Seq technology linked the cytotoxicity to cellular stress pathways, including oxidative stress and unfolded protein response (UPR). Induction of these stress pathways in response to amiodarone, diclofenac, and ibuprofen, was demonstrated to be concentration and time dependent. The transcriptomics data on diclofenac-treated HLCs were found to be more sensitive in detecting differentially expressed genes in response to treatment, as compared to existing datasets of other diclofenac-treated in vitro hepatocyte models. Hence iPSC-HLCs generated by transcription factor overexpression and in metabolically optimised medium appear suitable for chemical toxicity detection as well as mechanistic toxicity studies.

Chromosome-Level Genome Assembly Provides New Insights into Genome Evolution and Tuberous Root Formation of Potentilla anserina

Potentilla anserina is a perennial stoloniferous plant with edible tuberous roots in Rosaceae, served as important food and medicine sources for Tibetans in the Qinghai-Tibetan Plateau (QTP), China, over thousands of years. However, a lack of genome information hindered the genetic study. Here, we presented a chromosome-level genome assembly using single-molecule long-read sequencing, and the Hi-C technique. The assembled genome was 454.28 Mb, containing 14 chromosomes, with contig N50 of 2.14 Mb. A total of 46,495 protein-coding genes, 169.74 Mb repeat regions, and 31.76 Kb non-coding RNA were predicted. P. anserina diverged from Potentilla micrantha ∼28.52 million years ago (Mya). Furthermore, P. anserina underwent a recent tetraploidization ∼6.4 Mya. The species-specific genes were enriched in Starch and sucrose metabolism and Galactose metabolism pathways. We identified the sub-genome structures of P. anserina, with A sub-genome was larger than B sub-genome and closer to P. micrantha phylogenetically. Despite lacking significant genome-wide expression dominance, the A sub-genome had higher homoeologous gene expression in shoot apical meristem, flower and tuberous root. The resistance genes was contracted in P. anserina genome. Key genes involved in starch biosynthesis were expanded and highly expressed in tuberous roots, which probably drives the tuber formation. The genomics and transcriptomics data generated in this study advance our understanding of the genomic landscape of P. anserina, and will accelerate genetic studies and breeding programs.

Small transcriptional differences lead to distinct NF-κB dynamics in quasi-identical cells

Transcription factor dynamics is fundamental to determine the activation of accurate transcriptional programs and yet is heterogeneous at single-cell level. The source of this dynamic variability is not completely understood. Here we focus on the nuclear factor κB (NF-κB), whose dynamics have been reported to cover a wide spectrum ranging from oscillatory to non-oscillatory. We show that clonal populations of immortalized fibroblasts derived from a single mouse embryo (that can hence be considered quasi-identical) display robustly distinct dynamics upon tumor necrosis α (TNF-α) stimulation. Combining transcriptomics, data-constrained mathematical modelling, and mRNA interference we show that small differences in the expression of genes belonging to the NF-κB regulatory circuit are predictive of the distinct responses to inflammatory stimuli observed among the clones. We propose that this transcriptional fine-tuning can be a general mechanism to produce cell subpopulations with distinct dynamic responses to stimuli within homogeneous cell populations.

CCST: Cell clustering for spatial transcriptomics data with graph neural network

Spatial transcriptomics data can provide high-throughput gene expression profiling and spatial structure of tissues simultaneously. An essential question of its initial analysis is cell clustering. However, most existing studies rely on only gene expression information and cannot utilize spatial information efficiently. Taking advantages of two recent technical development, spatial transcriptomics and graph neural network, we thus introduce CCST, Cell Clustering for Spatial Transcriptomics data with graph neural network, an unsupervised cell clustering method based on graph convolutional network to improve ab initio cell clustering and discovering of novel sub cell types based on curated cell category annotation. CCST is a general framework for dealing with various kinds of spatially resolved transcriptomics. With application to five in vitro and in vivo spatial datasets, we show that CCST outperforms other spatial cluster approaches on spatial transcriptomics datasets, and can clearly identify all four cell cycle phases from MERFISH data of cultured cells, and find novel functional sub cell types with different micro-environments from seqFISH+ data of brain, which are all validated experimentally, inspiring novel biological hypotheses about the underlying interactions among cell state, cell type and micro-environment.