A data analysis framework for biomedical big data: Application on mesoderm differentiation of human pluripotent stem cells

Multi-assignment clustering: Machine learning from a biological perspective

A common approach for analyzing large-scale molecular data is to cluster objects sharing similar characteristics. This assumes that genes with highly similar expression profiles are likely participating in a common molecular process. Biological systems are extremely complex and challenging to understand, with proteins having multiple functions that sometimes need to be activated or expressed in a time-dependent manner. Thus, the strategies applied for clustering of these molecules into groups are of key importance for translation of data to biologically interpretable findings. Here we implemented a multi-assignment clustering (MAsC) approach that allows molecules to be assigned to multiple clusters, rather than single ones as in commonly used clustering techniques. When applied to high-throughput transcriptomics data, MAsC increased power of the downstream pathway analysis and allowed identification of pathways with high biological relevance to the experimental setting and the biological systems studied. Multi-assignment clustering also reduced noise in the clustering partition by excluding genes with a low correlation to all of the resulting clusters. Together, these findings suggest that our methodology facilitates translation of large-scale molecular data into biological knowledge. The method is made available as an R package on GitLab (https://gitlab.com/wolftower/masc).

Human Pluripotent Stem Cell-Derived Hepatocytes Show Higher Transcriptional Correlation with Adult Liver Tissue than with Fetal Liver Tissue

Human pluripotent stem cell-derived hepatocytes (hPSC-HEP) display many properties of mature hepatocytes, including expression of important genes of the drug metabolizing machinery, glycogen storage, and production of multiple serum proteins. To this date, hPSC-HEP do not, however, fully recapitulate the complete functionality of in vivo mature hepatocytes. In this study, we applied versatile bioinformatic algorithms, including functional annotation and pathway enrichment analyses, transcription factor binding-site enrichment, and similarity and correlation analyses, to datasets collected from different stages during hPSC-HEP differentiation and compared these to developmental stages and tissues from fetal and adult human liver. Our results demonstrate a high level of similarity between the in vitro differentiation of hPSC-HEP and in vivo hepatogenesis. Importantly, the transcriptional correlation of hPSC-HEP with adult liver (AL) tissues was higher than with fetal liver (FL) tissues (0.83 and 0.70, respectively). Functional data revealed mature features of hPSC-HEP including cytochrome P450 enzymes activities and albumin secretion. Moreover, hPSC-HEP showed expression of many genes involved in drug absorption, distribution, metabolism, and excretion. Despite the high similarities observed, we identified differences of specific pathways and regulatory players by analyzing the gene expression between hPSC-HEP and AL. These findings will aid future intervention and improvement of in vitro hepatocyte differentiation protocol in order to generate hepatocytes displaying the complete functionality of mature hepatocytes. Finally, on the transcriptional level, our results show stronger correlation and higher similarity of hPSC-HEP to AL than to FL. In addition, potential targets for further functional improvement of hPSC-HEP were also identified.

Transcriptional sex and regional differences in paired human atrial and ventricular cardiac biopsies collected in vivo

Transcriptional studies of the human heart provide insight into physiological and pathophysiological mechanisms, essential for understanding the fundamental mechanisms of normal cardiac function and how they are altered by disease. To improve the understanding of why men and women may respond differently to the same therapeutic treatment it is crucial to learn more about sex-specific transcriptional differences. In this study the transcriptome of right atrium and left ventricle was compared across sex and regional location. Paired biopsies from five male and five female patients undergoing aortic valve replacement or coronary artery bypass grafting were included. Gene expression analysis identified 620 differentially expressed transcripts in atrial and ventricular tissue in men and 471 differentially expressed transcripts in women. In total 339 of these transcripts overlapped across sex but notably, 281 were unique in the male tissue and 162 in the female tissue, displaying marked sex differences in the transcriptional machinery. The transcriptional activity was significantly higher in atrias than in ventricles as 70% of the differentially expressed genes were upregulated in the atrial tissue. Furthermore, pathway- and functional annotation analyses performed on the differentially expressed genes showed enrichment for a more heterogeneous composition of biological processes in atrial compared with the ventricular tissue, and a dominance of differentially expressed genes associated with infection disease was observed. The results reported here provide increased insights about transcriptional differences between the cardiac atrium and ventricle but also reveal transcriptional differences in the human heart that can be attributed to sex.

A Systems-level Characterization of the Differentiation of Human Embryonic Stem Cells into Mesenchymal Stem Cells

Mesenchymal stem/stromal cells (MSCs) are self-renewing multipotent cells with regenerative, secretory and immunomodulatory capabilities that are beneficial for the treatment of various diseases. To avoid the issues that come with using tissue-derived MSCs in therapy, MSCs may be generated by the differentiation of human embryonic stems cells (hESCs) in culture. However, the changes that occur during the differentiation process have not been comprehensively characterized. Here, we combined transcriptome, proteome and phosphoproteome profiling to perform an in-depth, multi-omics study of the hESCs-to-MSCs differentiation process. Based on RNA-to-protein correlation, we determined a set of high confidence genes that are important to differentiation. Among the earliest and strongest induced proteins with extensive differential phosphorylation was AHNAK, which we hypothesized to be a defining factor in MSC biology. We observed two distinct expression waves of developmental HOX genes and an AGO2-to-AGO3 switch in gene silencing. Exploring the kinetic of noncoding ORFs during differentiation, we mapped new functions to well annotated long noncoding RNAs (CARMN, MALAT, NEAT1, LINC00152) as well as new candidates which we identified to be important to the differentiation process. Phosphoproteome analysis revealed ESC and MSC-specific phosphorylation motifs with PAK2 and RAF1 as top predicted upstream kinases in MSCs. Our data represent a rich systems-level resource on ESC-to-MSC differentiation that will be useful for the study of stem cell biology.