Bulk RNA-seq

• Replaced microarrays in the late 00’s and has been widely used since
• Average expression level for 1000’s of genes
• Useful for:
  • Compare tissues
  • Compare disease / healthy
• Insufficient for studying
  • heterogeneous systems
    • early development
    • complex tissues (brain)
  • Stochastic gene expression

scRNA-seq

• A new technology, first publication by Tang 2009
• Widespread from ~2014: new protocols and lower sequencing costs
• Measures the distribution of expression levels for each gene

scRNA-seq evolution

scRNA-seq allows to study new biological questions

• Cell type identification
• Heterogeneity of response
• Stochasticity of gene expression
• Reconstructing continous processeses from a single experiment
• Holistic view, hypothesis-free science
  • Datasets range from \(10^2\) to \(10^6\) cells and increase in size every year
  • Not testing a concrete hypothesis
• Hypothesis-driven and data-driven science
  • Heard about it? Could you explain?

Hypothesis-driven and data-driven science

Basic Workflow

from Wikipedia

Cell capture

Manual isolation by mouth pipet

Laser capture micro-dissection (LCM) into eppendorf tubes

FACS sorting into microwell plates

{width=600px}

Isolation by FACS machine

• Fluorescence activated cell sorting
• You can select a subset of cells:
  • Size of blood cells
  • Dazl-GFP mouse line to select germ cells

FACS concept

Isolation by microfluidics & preparation insider droplets

> Macosko 2015

Picowell

Picocells allow simple sedimentation instead of FACS

• Too small for FACS
• Volume per well is so low, that even at 10% loading, you waste very little reagents

Platforms & scaling

Source: Svensson 2018

RNA amplification strategies

• Full length
• 3’ (aka: 3 prime)
• 5’ (aka: 5 prime)

RNA structure

Expressed sequence tag counting (EST)

Different Coverage along the transcript (Full L. vs. 3’)

Bhargava 2014

Full length

• Advantage?
  • Isoform specificity
  • Higher sensitivity
• SMART-seq2 [@Picelli2013-sb]

3’

• Sufficient for transcript counting
• Advantage?
  • A lot cheaper
• 3’ alternative polyadenylation
  • One kind of isoforms
• CEL-seq (Hashimshony 2012)
• Drop-seq (Macosko 2015)
• Almost all novel methods

5’

• Sufficient for transcript counting & cheap
• 5’: Alternative transcription start site
• STRT-seq (Islam et al.)

Steps in single-cell mRNA sequencing

Problems

• RNA is unstable ← → RNAse enzymes are very robust
  • Convert it to cDNA immediately
• Very low RNA input material → Huge amplification
• Huge amplification → Amplification bias (“PCR artefact”)

Reverse transcriptase

A reverse transcriptase (RT) is an enzyme used to generate complementary DNA (cDNA) from an RNA template, a process termed reverse transcription. Reverse transcriptases are used by:

• retroviruses to replicate their genomes,
• retrotransposon mobile genetic elements to proliferate within the host genome,
• eukaryotic cells to extend the [telomeres], and by
• some non-retroviruses such as the hepatitis B virus, a member of the Hepadnaviridae.

Wikipedia, edited

Reverse transcription creates uniquely tagged cDNA molecules

Velten 2015

Primer design & labelling strategy

Unique Molecular Identifiers (UMIs),

Kivioja 2012

In-vitro transcription or PCR amplifies your

Velten 2015

Linearly amplification by IVT is less noise sensitive to fluctuation in input material

Kolodziejczyk 2015

• After the invention of UMI, it only spares sequencing costs

Information in paired end reads

Velten 2015

Overview

{width=300px}

Velten 2015