+ All Categories
Home > Documents > Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short...

Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short...

Date post: 19-Aug-2020
Category:
Upload: others
View: 1 times
Download: 0 times
Share this document with a friend
28
Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory Amsterdam UMC [email protected] www.bioinformaticslaboratory.nl
Transcript
Page 1: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Single-cell RNA-Seq data: a (very) short overview

March 13, 2020

Dr. ir. Perry D. MoerlandBioinformatics LaboratoryAmsterdam [email protected]

Page 2: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

RNA-Seq

Gene expression => Quite easy (count the reads)Transcript expression => More difficult (but can be done probabilistically)Gene fusion => More difficult (esp. for new fusion events)Splicing => More difficult (esp. for poorly annotated isoforms)

RNA fragmentation

Adaptors + amplification

Sequencing

Map to reference transcriptomeRNA

ACCTAG…CGGTAA…ATGGCA…TGGGAC…TATAGG…

Reverse transcription

Gene A Gene B

Millions of reads

Measured expression is the average over all cell types

Page 3: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Single-cell RNA-Seq: motivation

Bulk sample gene expression analysis

We have information about all cell types

The signal is mixed with other cell types

Possible in some cases to infer the fraction of the different cell types, but (almost) impossible to infer their actual gene expression profile

Different expression in different cell types in mixed sample

Bimodal expression in a pure sample

Exam

ples

Same average

Same average

gene1 gene1

T1 T2 T1 T2

Page 4: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Influence of heterogeneity

Signal dilution false negatives

High proportiontumour cells

Low proportiontumour cells

Heterogeneous composition Associated with phenotype

Signal inflation false positives

1

0

0

1

0

0.90.1

Beta value

B

100 1

10 1

00

1 101 1

A

10 1

1

Condition Overall measured signal

Beta value of one probe (0 ≤ B ≤ 1)

0

Page 5: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Tang et al., Nature Methods (2009)

2009: n=1

Page 6: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Technological breakthroughs

DIY

Page 7: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Exponential scaling of single-cell RNA-seq in the past decade

Svensson et al., Nature Methods (2018)

2019Cao et al., Nature

Combinatorial indexing

Page 8: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

scRNA-Seq: workflow

https://en.wikipedia.org/wiki/Single_cell_sequencing

Page 9: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

scRNA-Seq: what can go wrong?

https://en.wikipedia.org/wiki/Single_cell_sequencing

Biased selection of cellsDuplicates

(doublets/multiplets)Broken/dead cells

Dropouts (60-90%)?

Amplification bias

Sequencing errorsRodriques et al., Science (2019): Slide-seq

Page 10: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Collect RTPool & Remove oil

OilCellsEnzymes

10x BarcodedGel Beads

Single CellGEMs

10x Barcoded

cDNA

10x BarcodedcDNA

Sequencing

Final Library Construct

10x Barcode

read1 read2

P5P7

UMI

Sample Index

Poly(dT)NN

Page 11: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Typical single-cell RNA-seq analysis workflow

Luecken et al., Molecular Systems Biology (2019)

Page 12: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Amplification bias: unique molecular identifiers (I)

http://mccarrolllab.com/dropseq/

Page 13: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

QC at cell level

Stegle et al., Nature Reviews Genetics (2015)

Page 14: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

QC at cell level (II)

Luecken et al., Molecular Systems Biology (2019)

Page 15: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

QC at gene level

Lun et al., F1000Research (2016)

Percentage of total counts assigned to the top 50 most highly-abundant features

Does it match expected biology?

Page 16: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Variability in scRNA-Seq data (I)

Kolodziejczyk et al., Molecular Cell (2015)

Noise

Signal

Extrinsic noise (regulation by transcription factors)vs intrinsIc noise (stochastic bursting/firing, cell cycle)

Overdispersion, batch effects, sequencing depth, GC bias, amplification bias

capture efficiency (starting amount of RNA)

Variability in bulk RNA-SeqAdditional variability in scRNA-seq

Page 17: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Variability in scRNA-Seq data (II)

Karchenko et al., Nature Methods (2014)

Cell cycle

Page 18: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

NormalizationMethods developed for bulk samples commonly used, but poor fit for scRNA-Seq data…

Page 19: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Spike-ins (I)

Addition of external controls ERCC spike-ins most widely used, mix consists of 92 mRNAs at

different concentrations Important to add equal amounts to each cell preferably in the

lysis buffer

Page 20: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Spike-ins (II)

Can be used to model Technical noise Drop-out rates Starting amount of RNA in the cell Data normalization

Two cells from a homogeneous population but with different total mRNA content

Page 21: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Amplication bias: unique molecular identifiers (II)

Page 22: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

scRNA-Seq: common applications

Page 24: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

HashTag Oligonucleotide

Stoekius et al. Gen. Biol., 2018

Page 25: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

https://www.biolegend.com/en-us/totalseq

Antibody DerivedTags

Page 26: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Demultiplexing HTO

Cells are assigned to belong to a certain HTO (or combination!) based on the sequenced HTO-tags (and cut-offs used…)

Some cells do contain multiple HTO’sMost of these contain HTO’s belonging to the

same subject, i.e. HTO_9/HTO_10 etc. -> Doublets

Page 27: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

HTO tag used tocluster

Page 28: Single-cell RNA-Seq data: a (very) short overview · Single-cell RNA-Seq data: a (very) short overview March 13, 2020 Dr. ir. Perry D. Moerland Bioinformatics Laboratory. Amsterdam

Antibody Derived Tags (ADTs)


Recommended