Microsoft Word - c2c12v17_supp.docx - 13059_2021_2505_MOESM1

Fig.

1: LR

Split

seq preprocessing, QC, and additional analysis. a

, Schematic

diagram of LR

Split

seq demultiplexing strategy.

, UMI per ranked barcode plots

before and after barcode correction (both axes log scaled).

, Median number of UMIs

per cell binned by reads per cell before and after barcode correction.

, Unfiltered

isoforms per novelty category across all cells in LR

Split

seq data.

, Gene lengths of

annotated genes detected in bulk only, single

cell only, an

d single

nucleus only (log

scale).

, Transcript lengths of annotated transcripts detected in bulk only, single

cell

only, and single

nucleus only (log scale).

, Distribution of number of exons in bulk long

reads (yellow), single

cell long reads (pink), a

nd single

nucleus long reads (blue).

Upset plot of novel in catalog (NIC) transcripts that passed filtering found in bulk data

compared to single cell data across all samples. Bars on the left indicate set size,

circles indicate various combinations of

samples, and bars on top indicate the number

of genes found in each combination. Outline colors indicate technology (bulk in yellow,

single

cell in magenta) and fill colors indicate sample type (72hr nuclei in green, 0hr

nuclei in blue, and 0hr cells in pi

nk for single

cell data; 72hr in green, 0hr in pink for bulk

data). Box plots above indicate gene length distribution for each intersection. Venn

diagrams below summarize the overlaps between bulk (left) and single

cell or single

nucleus (right), for each

sample type. Sample type is indicated by outline color.

, Upset

plot and Venn diagrams of novel not in catalog (NNC) transcripts that passed filtering

found in bulk data and single

cell data.

Fig.

2: Short

read Split

seq QC. a

, Schematic of sample t

ype per well in the first

round of barcoding (pink = 0hr cells, blue = 0hr nuclei, and green = 72hr nuclei). Panels

to the right show the number of cells per well across each round of barcoding for a

9,000

cell sublibrary.

, UMI per cell knee plots for th

e 1,000

cell sublibrary sequenced

with both long and short reads indicating a threshold of 3,936 reads per cell, leaving 568

cells before additional QC.

, Violin plots of scRNA

seq QC metrics after filtering for the

464 cells only.

, An example knee plot

for a 9,000

cell sublibrary indicating a threshold

of 370 reads per cell, leaving 7,405 cells before additional QC.

, Violin plots of scRNA

seq QC metrics after filtering for all cells.

Fig.

3: Short

read

and LR

Split

seq additional analysis. a

, Dis

tribution of marker

genes within the 464

cell UMAP (dark blue = lowly expressed, yellow = highly

expressed).

, Proportion of spliced vs. unspliced reads per cell in

short

read Split

seq

and LR

Split

seq

from RNA velocity analysis.

Cells are labeled by sam

ple type (0hr cells

in pink, 0hr nuclei in blue,

and 72hr nuclei in green) and marginals on the top and right

indicate their distributions

, Short

read (left) and LR

Split

seq (right) UMAPs for 464

cells with RNA velocity field trajectories indicated by arrows.

, Isoform complexity

(Number of genes with more than one isoform per cell) vs. number of reads per cell,

colored by sample type (0h

r cells in pink, 0hr nuclei in blue, and 72hr nuclei in green).

Fig.

4: Swan analysis of

Tpm2

and

Pkm

isoforms.

, Gene report made by Swan

for

Tpm2

. Relative expression of each isoform, separated by 0hr MB cells, 72hr

Pax7

nuclei, and 72hr

Myog

uclei plotted alongside the isoform’s name, transcript novelty,

and structure. Exons 6a and 6b, known to be alternatively spliced during C2C12

differentiation, are highlighted.

, Gene report made by Swan for

Pkm

, separated by the

same cell types. Mutually

exclusive exons 9 and 10 as well as alternative TES in Pkm

205 are highlighted.

Fig.

5: Additional analysis of 38,000

cell short

read Split

seq data.

, UMAP of

464 cells with both short and long reads colored by 20 clusters derived using 36,869

short

read cells.

, Heatmap of cell cycle marker genes in the 20 clusters.

Visualization of

Myog

in mononucleated cells and myotubes at the 72hr differentiat

ion

timepoint. Blue = DAPI, green =

Myog

. Scale bar: 50 μm.

, Heatmap of marker genes

in the 20 clusters (dark blue = low expression, yellow = high expression).

, Dot plot of

transcription factors and marker genes involved in myogenesis found from differ

ential

expression testing and/or literature. Genes that did not pass the differential expression

threshold yet are of interest in the system and significantly expressed in prior classic

bulk data are colored grey (

Id1

Id2

Myod1

Myf5

Tcf3,

and

Tcf12

Fig.

6: Additional analysis/QC of snATAC

seq. a

, UMI per barcode knee plot for an

example snATAC

seq library (0hr, 6,782 nuclei).

, Violin plots of snATAC

seq QC

metrics after filtering > 6 TSS enrichment, < 20,000 reads, and > 5,000 reads per

nucleus.

, Distribution of marker genes within the UMAP colored by gene activity score

(dark blue = low activity, yellow = high activity).

, Integration of scRNA

seq and

snATAC

seq data, labeled by cell type (0hr in pink and 72hr in green on left; MB in

pink,

Myog

in dark green, and

Pax7

in light green on right).

, Pseudobulk peaks per

cluster spanning the

Pax7

locus. TSS track indicates TSSs called from LR

Split

seq

data.

, Pseudobulk peaks spanning the

Myog

and

Mybph

loci.

, Heatmap of top 50

rker regions in the 18 snATAC

seq clusters (dark blue = low accessibility, yellow =

high accessibility).

, Cluster A9 GO term enrichment clustergram. Examples of genes

associated with A9 marker peaks belonging to the GO terms in rows are indicated in

red.

Fig.

7: Identification and validation of TSSs/TESs from long

read data. a

Histogram of number of LR

Split

seq reads supporting each TSS.

, Bubble plot of the

number of distinct exon combinations (splice isoforms) detected per gene compared to

the nu

mber of distinct TSSs detected per gene in bulk data.

, Validation of TSSs found

in bulk long

read data using 4 external datasets (ENCODE proximal enhancer and

promoter cCREs, GENCODE TSSs, and CAGE peaks) and our snATAC

seq

pseudobulk peaks.

, Bubble pl

ot of the number of distinct exon combinations (splice

isoforms) detected per gene compared to the number of distinct TESs detected per

gene found in long

read bulk data.

, Validation of TESs found in bulk long

reads using

GENCODE TESs and polyA

seq data.

, Bubble plot of splice isoforms per gene per cell

compared to TESs detected per gene per cell found in LR

Split

seq.

, Validation of