spatialLIBD webinar

29
Transcriptome-Scale
Spatial Gene Expression in the
Human Dorsolateral Prefrontal Cortex
Kristen R Maynard, Ph.D., Research Scientist
Leonardo Collado-Torres, Ph.D., Research Scientist
Lieber Institute for Brain Development
TheScientist Webinar
March 19, 2020
@kr_maynard
@fellgernon
@LieberInstitute
@TheScientistLLC

View Slide

The spatial architecture of the brain is
fundamentally connected to its function
2
chartdiagram.com slideshare.net

View Slide

Laminar position of a cell influences its gene
expression, morphology, physiology, and
function
3
Kwan et al., 2012, Development

View Slide

Single nucleus RNA-sequencing & Visium technologies
4
Single Cell Gene Expression
Spatial Gene Expression

View Slide

Webinar Overview
5
1. Identification of layer-enriched genes in human cortex using Visium.
2. Spatial registration of single-nucleus RNA-seq data from human cortex.
3. Layer-enriched expression of genes associated with brain disorders.
Maynard, Collado-Torres, et al, bioRxiv, 2020

View Slide

Study design for Visium experiments in
dorsolateral prefrontal cortex (DLPFC)
6
Andrew E Jaffe
Keri Martinowich
Stephanie C Hicks Lukas M Weber
Cedric Uytingco Nikhil Rao
@stephaniehicks @lmwebr @martinowk @andrewejaffe

View Slide

Visualizing gene expression in a histological context
7
logcounts logcounts logcounts

View Slide

2 pairs spatial adjacent replicates x subject = 12 sections
8
Subject 1
Subject 2
Subject 3
Adjacent spatial replicates (0µm) Adjacent spatial replicates (300µm)
PCP4

View Slide

“Pseudo-bulking” collapses data: spot to layer level
9

View Slide

Three statistical models to assess laminar enrichment
“ANOVA”
model
10
“Enrichment”
model
“Pairwise”
model
Is any layer different? Is one layer > the rest? Is layer X > layer Y?

View Slide

11 ISH images courtesy of Allen Human Brain Atlas: http://human.brain-map.org/ (Hawrylycz et al., 2012)
Visium replicates layer-enrichment of previously
identified layer marker genes
L4>rest, p=1.74e-09
L6>WM, p=4.48e-19
logcounts
logcounts

View Slide

Identification & validation of novel layer-enriched genes
12
L5>rest, p=4.33e-12
L6>rest, p=5.05e-12
L1>rest, p=1.47e-10
L2>rest, p=9.73e-11

View Slide

Visium layer-enriched vs. canonical marker genes
13
L1>rest, p=7.94e-15 L5>L3, p=4.44e-02

View Slide

14
Segmentation of histology data identifies spots
containing single cell bodies and neuropil
50um
Gray matter White matter
Neuron
Neuropil
Glial cell
Mouse Brain Tissue Postmortem Human DLPFC
Madhavi Tippani
@MadhaviTippani
Joseph L Catallini II

View Slide

L4
L3
L2
L1

(A) (B)
(C)
Spatial registration of your sc/snRNA-seq data
Your sc/snRNA-seq data
Hodge et al, Nature, 2019

View Slide

L4
L3
L2
L1

(A) (B)
(C)
Your sc/snRNA-seq data
Our spatial data
Hodge et al, Nature, 2019

View Slide

17
12
17 Matthew N Tran Brianna K Barry
@mattntran @sudo_BreeB
Identify clusters in your sc/snRNA-seq data
- Pre-process your sc/snRNA-seq data
- Identify cell/nuclei clusters
- Find data-driven marker genes and/or
combine with known marker genes
- Label clusters

View Slide

18
# columns for us:
12 * 7 = 84 (76)
“Pseudo-bulk” our spatial transcriptomics data

View Slide

19
Your sc/snRNA-seq:
cell or nuclei clusters * subjects or other analysis variables
“Pseudo-bulk” your sc/snRNA-seq data

View Slide

Three statistical models to assess laminar enrichment
“ANOVA”
model
20
“Enrichment”
model
“Pairwise”
model
Is any layer different?
Is one layer > the rest?
Is layer X > layer Y?

View Slide

WM
L6
L5
L4
L3
L2
L1
Oli3
Oli5
Oli4
Oli0
Oli1
Ast3
Ast2
Ast0
Ast1
Mic2
Mic3
Mic0
Mic1
Opc0
Opc1
Opc2
Per
End1
End2
Ex2
Ex0
Ex4
Ex6
Ex14
Ex1
Ex5
Ex7
Ex8
In0
In7
In9
In11
In2
In10
In3
In6
In1
In4
In5
In8
Ex3
Ex11
Ex12
Ex9
í
í
í
í

(C)
Interpretation guidelines:
• Find strong positive correlation values (dark green) to identify
cell/nuclei clusters enriched for a given layer
• By row: for a given layer
• By column: for a given cell/nuclei cluster
Mathys et al, Nature, 2019

View Slide

WM
Layer6
Layer5
Layer4
Layer3
Layer2
Layer1
22 (Oligo)
3 (Oligo)
23 (Oligo)
17 (Oligo)
21 (Oligo)
7 (Astro)
5 (Astro)
9 (OPC)
26 (OPC)
1 (Micro)
24 (Drop)
13 (Excit)
10 (Excit)
27 (Excit)
29 (Inhib)
14 (Inhib)
15 (Inhib)
18 (Inhib)
2 (Excit)
31 (Excit)
8 (Excit)
16 (Inhib)
28 (Inhib)
30 (Inhib)
20 (Inhib)
11 (Inhib)
25 (Inhib)
4 (Excit)
12 (Excit)
6 (Excit)
19 (Excit)
−0.8
−0.6
−0.4
−0.2
0.0
0.2
0.4
0.6
0.8
Matthew N Tran Brianna K Barry
@mattntran @sudo_BreeB
Interpretation guidelines:
• Find strong positive correlation values (dark green)
to identify cell/nuclei clusters enriched for a given
layer
• By row: for a given layer
• By column: for a given cell/nuclei cluster

View Slide

http://spatial.libd.org/spatialLIBD/
Spatial registration of your sc/snRNA-seq data: DIY

View Slide

24
Cluster1 Cluster2 Cluster3
ENSG00000104419 3 -2 0.3
ENSG0000018400
7
1 0.67 4
… … … …
Full example table
https://github.com/LieberInstitute/spatialLIBD/blob/master/data-raw/tstats_Human_DLPFC_snRNAseq_Nguyen_topLayer.csv
Save your “enrichment” t-
statistics for your sc/snRNA-seq
clusters
Spatial registration of your sc/snRNA-seq data: DIY

View Slide

25
Spatial registration of your sc/snRNA-seq
data: DIY
spatial.libd.org/spatialLIBD/
Cluster1 Cluster2 Cluster3
ENSG00000104419 3 -2 0.3
ENSG00000184007 1 0.67 4
… … … …

View Slide

Gandal et al, Science, 2018
SFARI GENE; 2.0 by Abrahams et al, Mol Autism, 2013
Jaffe et al, Nature Neuroscience, 2020
- Curated lists
- GWAS/TWAS
hits
- Differential
expression
- …
Layer-enriched gene expression profiling

View Slide

0
2
4
6
8
10
12
WM
L6
L5
L4
L3
L2
L1
SFAR
I
ASC
102
ASD
53
D
D
ID
49
D
E.U
p
D
E.D
ow
n
2.7
2.1
2.7
4
3.6
4.9
4.5
2.5
5
2.8
5
6.4
2.8
ASD
0
2
4
6
8
10
12
WM
L6
L5
L4
L3
L2
L1
PE.U
p
PE.D
ow
n
BS2.U
p
BS2.D
ow
n
BS2.U
p
BS2.D
ow
n
PE.U
p
PE.D
ow
n
2.1
2
3.1
1.8
2.2
1.8
8.8
5
2.7
2.6
4.6
6&='í'( 6&='í7:$6
(A) (B)
DIY at
Alzheimer’s Disease
• SFARI: Abrahams et al, Mol Autism,
2013
• ASC102: Satterstrom et al, Cell,
2020
Break up into:
• ASD53: ASD dominant traits
• DDID49: neurodevelopmental
delay

View Slide

0
2
4
6
8
10
12
WM
L6
L5
L4
L3
L2
L1
SFAR
I
ASC
102
ASD
53
D
D
ID
49
D
E.U
p
D
E.D
ow
n
2.7
2.1
2.7
4
3.6
4.9
4.5
2.5
5
2.8
5
6.4
2.8
ASD
0
2
4
6
8
10
12
WM
L6
L5
L4
L3
L2
L1
PE.U
p
PE.D
ow
n
BS2.U
p
BS2.D
ow
n
BS2.U
p
BS2.D
ow
n
PE.U
p
PE.D
ow
n
2.1
2
3.1
1.8
2.2
1.8
8.8
5
2.7
2.6
4.6
6&='í'( 6&='í7:$6
(A) (B)
DIY at
Gandal et al, Science, 2018
Collado-Torres et al, Neuron, 2019

View Slide

29
Stephanie C Hicks Lukas M Weber
@stephaniehicks @lmwebr Maynard, Collado-Torres, et al, bioRxiv, 2020
Data-driven layer-enriched clustering in the DLPFC
Spatially-varying genes
Highly-variable genes
Spot-level
clustering
Manual layer
annotation
using
spatialLIBD
• Which samples to use?
• All samples?
• Sample by sample then
merge?
• Use image-derived information?

View Slide

30
SpatialDE by Svensson et al, Nature Methods, 2018
Are the spatial patterns relevant?
Remember to inspect your data!

View Slide

31
SpatialDE by
Svensson et al, Nature Methods, 2018
“ANOVA” model F-statistics
SpatialDE
statistic

View Slide

32
Use known
marker
genes only
Use layer-
enriched genes
(scenario where
you have more
datasets)
Only use
the data
Requires >=1 expert
Benefits from known marker
genes (if expressed) & prior
knowledge

View Slide

33
Using spatial coordinates does help in some cases

View Slide

Explore our spatial data (or adapt for yours) + perform
spatial registration & gene enrichment analyses

View Slide

Summary: transcriptome-scale spatial gene
expression in postmortem human cortex
35
http://research.libd.org/spatialLIBD
Explore the data:

View Slide

Acknowledgements
Lieber Institute
Keri Martinowich
Andrew E. Jaffe
Brianna K. Barry
Joseph L. Catallini II
Matthew N. Tran
Zachary Besich
Madhavi Tippani
Joel E. Kleinman
Thomas M. Hyde
Daniel R. Weinberger
JHU Biostatics Dept JHU Oncology Tissue Services (Kristen Lecksell)
Stephanie C. Hicks JHU SKCCC Flow Core (Jessica Gucwa)
Lukas M. Weber JHU Transcriptomics & Deep Sequencing Core (Linda Orzolek)
10x Genomics
Cedric Uytingco
Stephen R. Williams
Jennifer Chew
Yifeng Yin
Nikhil Rao
36
@kr_maynard
@fellgernon
@LieberInstitute
@TheScientistLLC
Interested in working with us?
Let us know!

View Slide

spatialLIBD webinar

spatialLIBD webinar

Leonardo Collado-Torres

More Decks by Leonardo Collado-Torres

Other Decks in Science

Featured

Transcript