Posters - Schedules

Posters Home

Uterine fibroids (leiomyomas) are clonal tumors originating from the myometrium (Myo), affecting over 70% of women of reproductive age. While abnormal epigenetic modifications and transcription factor chromatin binding have been implicated in developing leiomyomas (Meio) with mutations in MED12, understanding their impact on gene expression across diverse regulatory mechanisms is challenging. This work extends our established deep convolutional neural network (iSEGnet) to investigate the relationship between epigenetic modification (H3K27ac), 5 transcriptional regulators, and gene expression regulation in MED12 mutant uterine leiomyomas. Using published omics profiles of 5 Leio and Myo tissues, we show that the two iSEGnet models (Leio and Myo, respectively) had moderate capabilities for predicting gene expression, with R2 values of 0.42 and 0.38 and Spearman-correlation values of 0.62 and 0.6 between the predicted and observed expression values, respectively. Using DeepLiftShap we further evaluate the attribution score linked to a particular ChIP-seq profile at DNA base pair position. We show that among the top genes ranked by the maximum difference of the Shapely scores between Myo and Leio, 8 collagen genes had strongest predictive influence from RNAPII, and other 6 AP-1 subunit members had either CDK8 or RNAPII as their most predictive factor.

Genome-wide identification of single-cell transcriptomic responses of drug candidate compounds in various human cells is a challenging issue in medical and pharmaceutical research. In this study, we present a computational method, tensor-based imputation of gene-expression data at the single-cell level (TIGERS), in order to reveal the drug-induced single-cell transcriptomic landscape. With this algorithm, we predict missing drug-induced single-cell gene-expression data with tensor imputation, and identify trajectories of regulated pathways considering intercellular heterogeneity. Our proposed tensor imputation method outperformed existing imputation methods for single-cell data completion, and provided cell-type-specific transcriptomic responses for unobserved drugs. For example, TIGERS was able to correctly predict the cell-type-specific expression of maker genes for pancreatic islets. Pathway trajectory analysis of the imputed gene-expression profiles of all combinations of drugs and human cells successfully identified single-cell-specific drug activities and pathway trajectories that reflect drug-induced changes in pathway regulation. The proposed method is expected to be useful for understanding of the single-cell mechanisms of drugs at the pathway level.

Transposon-mediated introduction of transgenes into genomes is a convenient way to enable host organisms to express exogenous genes. This approach has been particularly useful for biomanufacturing of therapeutic antibodies, which are introduced into genome of Chinese hamster ovary (CHO) cells enabling large-scale production in bioreactors. To gain detailed understanding of the mechanisms underlying transgene expression, it is crucial to accurately quantify number of insertions and their respective positions within the genome. Still, many methods aimed at this are not revealing an extensive picture, for instance, overlook fragmented insertions or transgenes in heterochromatic regions. To address this problem, we employ ultra-long read nanopore sequencing of whole genomes from multiple CHO clones expressing transgenic antibodies and develop novel methods comprehensively characterizing transgenic insertions. We then highlight the advantages and drawbacks of the ultra-long-read approach by comparing to the results obtained through TLA-seq and ddPCR. Finally, we perform a multi-omics analysis associating identified transgene insertions with RNA-seq expression profiles, open chromatin regions from ATAC-seq, methylation marks to gain unprecedented insights into transposon-mediated transgene integrations and their effect on expression dynamics. Our work emphasizes the importance of leveraging ultra-long whole genome sequencing for characterizing transgenic insertions and demonstrates its significant role in multi-omics approaches.

Major advances in bio-based industry are primed to fuel the transition towards environmentally sustainable and cost-efficient production schemes. Mastering metabolism of gaseous feedstocks such as carbon dioxide can support a sustainable carbon-neutral economy. Gas fermentation using acetogenic microorganisms offers a solution of increasingly demonstrated value to produce biofuels and value-added biochemicals. In spite of their prominent biosynthetic potential, a systemic understanding of the translational regulation of the acetogenic metabolism remains unclear. Since protein synthesis in bacteria is highly energy consuming, and acetogens grow at the thermodynamic limit of life under autotrophy, translational regulation cannot be escaped. Clarification of gene expression at the translation level is useful for rationally conceiving genetic modifications of acetogens to increase the efficiency of inorganic carbon fixation and subsequent product formation. RNA sequencing and ribosome profiling provide relevant information about gene transcription and mRNA translation. We gathered publicly available RNA sequencing and ribosome profiling data of several acetogens cultivated under heterotrophic and autotrophic conditions, providing data on genome-scale transcriptional and translational responses of acetogens during CO2 fixation. We then integrated genome-scale transcriptomic and translatomic data to untangle the complex relationship between transcription and translation and, thus, to fill a significant knowledge gap in the genotype-to-phenotype relationships.

Genetic regulation of pre-mRNA splicing (sQTLs) is a fundamental mechanism that affects complex traits and diseases. Existing large-scale sQTL studies conducted with tissue-level RNA-seq cannot resolve the cellular heterogeneity of genetic regulatory mechanisms. We provide a detailed dissection of the cell-type specific genetic regulation of pre-mRNA splicing by single-cell RNA sequencing of 1,058,909 peripheral blood mononuclear cells from 503 individuals. We demonstrate robust splicing quantification and reproducible genetic effects on splicing using replicate samples. We identify thousands of independent cis-sQTLs and hundreds of trans-sQTLs, most of which have regulatory effects orthogonal to eQTLs. Furthermore, we discovered a substantial number of cell-type specific sQTLs across 19 immune cell subtypes, as well as sex- and ancestry-biased sQTLs for genes known to be involved in autoimmune diseases. We next identified the dynamic usage of introns and changes in sQTL effects across the developmental trajectory from naive to memory B cells. Finally, we observed strong enrichment of sQTL effects in autoimmune GWAS loci and applied complementary colocalization and transcriptome-wide association approaches to pinpoint hundreds of cell-type specific putative causal genes for autoimmune GWAS. This work highlights the feasibility and importance of cell-type specific sQTL and their involvement in complex autoimmune diseases.