SINCELLMOD-2020: SINGLE CELL DATA IN NETWORK MODELING
PROGRAM FOR TUESDAY, DECEMBER 1ST
Days:
next day
all days

View: session overviewtalk overview

13:50-14:00 Session 1
13:50
Introduction
14:00
Making optimal use of your single-cell transcriptomic data: Building an analysis pipeline

ABSTRACT. The promise of single-cell RNA-seq is attracting a growing number of computational scientists that use and develop analysis tools. As more analysis tools are becoming available, it is becoming increasingly difficult to navigate this landscape and produce an up-to-date workflow to analyse one’s data. In this talk, I describe the steps of a typical single-cell RNA-seq analysis pipeline, and show the effects that method choice can have on the analysis results. Furthermore, I present analysis best-practice recommendations that we formulated based on independent comparison studies. As these recommendations are not tied to a particular programming language, I will also showcase our solution for building pipelines that can incorporate both python and R. We have built our recommendations into a best-practices workflow (available at https://www.github.com/theislab/single-cell-tutorial), which can serve as a workflow tutorial for new entrants into the field, and help established users update their analysis pipelines.

15:00-15:30Cellenion sponsor presentation & Coffee break

cellenONE®: Single Cell Isolation & Nanoliter Dispensing for Omics ApplicationscellenONE® is a revolutionary automated single cell isolation and dispensing systemIt combines gentle piezo-acoustic technology and image-based cell selection to allow highly accurate single cell isolation and miniaturization of library preparation. This unique open platform has already enabled wide ranging applications in both single cell analyses (scRNA-Seq, ATAC-Seq, scWGA and scMS) and cell line development (mAb, Stem Cells, CRISPR/CAS9).

Speaker 

Dr Guilhem Tourniaire, Scientific and Managing Director, Cellenion FRANCE

15:30-17:10 Session 2

Selected Abstracts

15:30
Synthesis of Boolean Networks from Single Cell Trajectory-based Constraints
PRESENTER: Andrei Zinovyev

ABSTRACT. Boolean networks model finite discrete dynamical systems with complex behaviours. The state of each component is determined by a Boolean function of the state of (a subset of) the components of the network. My talk will address the synthesis of these Boolean functions from constraints on their domain and emerging dynamical properties of the resulting network. The dynamical properties relate to the existence and absence of single cell trajectories between partially observed configurations, and to the stable behaviours (fixpoints and cyclic attractors). The synthesis is expressed as a Boolean satisfiability problem relying on Answer-Set Programming with a parametrized complexity, and leads to a complete non-redundant characterization of the set of solutions. Considered constraints are particularly suited to address the synthesis of models of cellular differentiation processes, as illustrated on a case study.

15:55
Metastability and energetic landscape : from Gene Regulatory Networks to cell trajectories

ABSTRACT. Differentiation is the process whereby a cell acquires a specific phenotype, by differential gene expression as a function of time. We can represent this process as a trajectory in the gene expression space. A popular vision of this evolution, introduced by Waddington, is to compare cells to marbles following probabilistic trajectories, as they roll through a developmental landscape of ridges and valleys. The landscape is often regarded to be shaped by the underlying dynamics of a Gene Regulatory Network (GRN).\\ Single cell technologies have shown that differentiation appears as a highly stochastic process. A model of gene expression able to recreate the noisy nature of gene expression from the GRN was recently introduced in Herbach et al (2017). It uses the formalism of Piecewise Deterministic Markov Processes (PDMP), and its results correspond accurately to the non-Gaussian distribution of the single cell gene expression data. We develop a mathematical framework and a method to express the energetic landscape of cell differentiation processes when described by the PDMP model. Combining mathematical results of Bresslof et al (2017), and methods developed in lv et al, (2014), we go further into the concept of metastability, highlighting the importance of a certain potential function able to characterize optimal trajectories of switching and the energetic barrier between different metastable basins, which can be seen as cell types. This approach is promising as it gives robust tools for interpreting the concept of landscape deformation. Following a development of Freidlin and Wentzell (1998), we also reduce the PDMP model to a simpler discrete model on a small number of cell types, which allows to derive an approximation of the protein stationary distribution.

16:20
Machine Learning based Data Imputation for Single-Cell ChIP-seq

ABSTRACT. Single-cell ChIP-seq analysis is challenging due to data sparsity. We present SIMPA (https://github.com/salbrec/SIMPA), a single-cell ChIP-seq data imputation method leveraging predictive information within bulk ENCODE data to impute missing protein-DNA interacting regions of target histone marks or transcription factors. Machine learning models trained for each single cell, each target, and each genomic region enable drastic improvement in cell types clustering and genes identification.

16:45
A multivariate statistical framework for characterizing transcriptional heterogeneity of single-cell profiles from rare diseases

ABSTRACT. 1. Introduction Single-cell RNA-sequencing (scRNA-seq) offers a powerful approach to characterize cellular heterogeneity associated to cell differentiation processes including intermediate cell-state transitions. In a clinical context, cell heterogeneity profiled through single-cell RNA-seq may uncovered relevant cell types and/or cell states associated to disease onset and progression. Notwithstanding, comparative analyses across patients and control samples are challenged by (i) technical stochasticity and batch effects, and (ii) environmental and physiological factors including age, sex, life style and clinical history [1].

2. Results Here we present a comprehensive multivariate statistical framework for the single-cell transcriptome analysis of rare genetic diseases. First, Cell-ID, a method based on Multiple Correspondence Analysis, is applied to extract a cell identity card in the form of an unbiased per-cell gene signature for each individual cell in a dataset. Per-cell signatures, or cell fingerprints, allow: (i) automatic cell type prediction using reference cell type signatures, and (ii) functional enrichment analysis using gene sets representing functional ontologies and pathways. More interestingly for the study of rare diseases, Cell-ID is able to extract gene signatures from rare cell sub-populations (<2%), and subsequently to “blast” such signatures against query datasets, i.e.: to test for a statistically robust replication of the newly uncovered cell signatures across samples from different donor. We applied Cell-ID to the single-cell transcriptome analysis of human hematopoietic stem and progenitor cells (HSPC) from mobilized peripheral blood in healthy donors. Cell-ID identified the most immature hematopoietic stem cells (HSC) and the prototypical hematopoietic progenitors along the myeloid and lymphoid branches. Notably, the Cell-ID multi-class evaluation allowed the identification of all the intermediate cell-states transitioning from the HSC at the top of the hierarchy to the various committed progenitors of the hematopoietic lineage. We used the underlying gene expression patterns to reconstruct an unbiased cell-state network in healthy conditions that will serve as a reference control for the comparative analysis of samples from patients with hematological disorders. Second, Sample-ID, a complementary method of Cell-ID that relies on Principal Component Analysis, is applied to extract a sample identity card in the form of unbiased gene signatures characterizing the observed transcriptional heterogeneity within a sample. Per-sample signatures, or sample fingerprints, allow in turn to “blast” query patient samples against reference single-cell RNA-seq. Both Cell-ID and Sample-ID are being systematically applied to single-cell RNA-seq datasets from (i) the Human Cell Atlas project, profiling healthy human organs and tissues, and (ii) in-house collections of rare disease patients profiling the affected organs and tissues.

3. References Oliver Stegle, Sarah A Teichmann and John C Marioni. Computational and analytical challenges in single-cell transcriptomics. Nat Rev Genet. 2015 Mar;16(3):133-45.