Tutorials

In this section we present several tutorials on using SCENIC+.

Basic SCENIC+ analysis

From count matrix and fragments file to enhancer driven gene regulatory network (eGRN)

For getting started with SCENIC+, we recommend this tutorial covering the entire SCENIC+ workflow including preprocessing steps. For this tutorial we make use of the 3k PBMCs multiome dataset publicly available from 10x Genomics. We’ll cover scRNA-seq and scATAC-seq preprocessing, topic modelling, motif enrichment analysis, running SCENIC+ and basic downstream analysis.

Get started by clicking here 👈

_images/B_cell_network.png

Running SCENIC+ step by step

In the PBMC tutorial (above) we used a wrapper function to run the complete SCENIC+ workflow. This workflow actually contains many individual steps. If you’re curious for these steps or you want to alter the workflow then this tutorial is for you. Here, we will showcase SCENIC+ on multiome data from the human cerebellum. We don’t show the preprocessing steps but we really focus on the individual steps of SCENIC+

Get started by clicking here 👈

_images/cerebellum_network.png

SCENIC+ analysis on non-multiome data: separate scATAC-seq and scRNA-seq on different cells of the same sample

It is possible to run the SCENIC+ workflow on samples for which you have separate scATAC-seq and scRNA-seq on different cells of the same samples. Here, we will generate pseudo-multiome data by sampling cells from the scRNA-seq and scATAC-seq experiment and combining them into metacells. This sampling happens within the same celltype. So one prerequisite to be able to run the analysis is a good annotation of the scATAC-seq data which matches the scRNA-seq annotation, this can be quite challenging. In this example we will analyze a mix melanoma cell lines (probably one of the few examples were the annotation is easy).

Get started by clicking here 👈

_images/mm_lines_dotplot.png

Advanced downstream analysis

Transcription factor perturbation simulation

The predicitons of SCENIC+ can be used to simulate the effect of transcription factor perturbations. We will illustrate this in the melanoma cell line analysis.

Get started by clicking here 👈

_images/mm_SOX10_kd_sim.png

Unbranched GRN velocity

The predictions of SCENIC+ can be used to predict in which differentiation “direction” a certain TF will drive a certain cell. This is term which we call GRN velocity (inspired by RNA velocity). These predictions make use between the lag in the sequence of events between TF expression, region accessibility and target gene expression. In this tutorial this principle is showcased along an unbranched differentiation trajectory, that of oligodendrocyte differentiation.

Get started by clicking here 👈

_images/unbranched_velocity.png

Branched GRN velocity

The predictions of SCENIC+ can be used to predict in which differentiation “direction” a certain TF will drive a certain cell. This is term which we call GRN velocity (inspired by RNA velocity). These predictions make use between the lag in the sequence of events between TF expression, region accessibility and target gene expression. In this tutorial this principle is showcased along a branched differentiation trajectory, that of the eye-antennal disk development in fly.

Get started by clicking here 👈

_images/branched_velocity.png