Integrative genomics and transcriptomics analysis of human embryonic and induced pluripotent stem cells

Combined ultra-low input mRNA and whole-genome sequencing of human embryonic stem cells

Background

Next Generation Sequencing has proven to be an exceptionally powerful tool in the field of genomics and transcriptomics. With recent development it is nowadays possible to analyze ultra-low input sample material down to single cells. Nevertheless, investigating such sample material often limits the analysis to either the genome or transcriptome. We describe here a combined analysis of both types of nucleic acids from the same sample material.

Methods

The method described enables the combined preparation of amplified cDNA as well as amplified whole-genome DNA from an ultra-low input sample material derived from a sub-colony of in-vitro cultivated human embryonic stem cells. cDNA is prepared by the application of oligo-dT coupled magnetic beads for mRNA capture, first strand synthesis and 3’-tailing followed by PCR. Whole-genome amplified DNA is prepared by Phi29 mediated amplification. Illumina sequencing is applied to short fragment libraries prepared from the amplified samples.

Results

We developed a protocol which enables the combined analysis of the genome as well as the transcriptome by Next Generation Sequencing from ultra-low input samples. The protocol was evaluated by sequencing sub-colony structures from human embryonic stem cells containing 150 to 200 cells. The method can be adapted to any available sequencing system.

Conclusions

To our knowledge, this is the first report where sub-colonies of human embryonic stem cells have been analyzed both at the genomic as well as transcriptome level. The method of this proof of concept study may find useful practical applications for cases where only a limited number of cells are available, e.g. for tissues samples from biopsies, tumor spheres, circulating tumor cells and cells from early embryonic development. The results we present demonstrate that a combined analysis of genomic DNA and messenger RNA from ultra-low input samples is feasible and can readily be applied to other cellular systems with limited material available.

Supervised learning methods in modeling of CD4+ T cell heterogeneity

Background

Modeling of the immune system – a highly non-linear and complex system – requires practical and efficient data analytic approaches. The immune system is composed of heterogeneous cell populations and hundreds of cell types, such as neutrophils, eosinophils, macrophages, dendritic cells, T cells, and B cells. Each cell type is highly diverse and can be further differentiated into subsets with unique and overlapping functions. For example, CD4+ T cells can be differentiated into Th1, Th2, Th17, Th9, Th22, Treg, Tfh, as well as Tr1. Each subset plays different roles in the immune system. To study molecular mechanisms of cell differentiation, computational systems biology approaches can be used to represent these processes; however, the latter often requires building complex intracellular signaling models with a large number of equations to accurately represent intracellular pathways and biochemical reactions. Furthermore, studying the immune system entails integration of complex processes which occur at different time and space scales.

Methods

This study presents and compares four supervised learning methods for modeling CD4+ T cell differentiation: Artificial Neural Networks (ANN), Random Forest (RF), Support Vector Machines (SVM), and Linear Regression (LR). Application of supervised learning methods could reduce the complexity of Ordinary Differential Equations (ODEs)-based intracellular models by only focusing on the input and output cytokine concentrations. In addition, this modeling framework can be efficiently integrated into multiscale models.

Results

Our results demonstrate that ANN and RF outperform the other two methods. Furthermore, ANN and RF have comparable performance when applied to in silico data with and without added noise. The trained models were also able to reproduce dynamic behavior when applied to experimental data; in four out of five cases, model predictions based on ANN and RF correctly predicted the outcome of the system. Finally, the running time of different methods was compared, which confirms that ANN is considerably faster than RF.

Conclusions

Using machine learning as opposed to ODE-based method reduces the computational complexity of the system and allows one to gain a deeper understanding of the complex interplay between the different related entities.