Transcript TAL1 OS enhancer tests, Figures

2593
2500
2000
1500
922
1000
596
292
500-2000
50-500
2-5
34
1-2
27
0.1-1
0
38
0.05-0.1
500
413
0-0.05
Number of TAL1 peaks
3000
5-50
C
Distance to the closest TSS (kb)
Figure S1: Genome-wide distribution of positions of TAL1 OSs relative to the transcription
start sites (TSSs) of RefSeq genes [110].
(A)
(B)
TAL1 peak
ID
VISTA
ID
Enhancer activity in
K562 (Fold change)
Tissues with enhancer activity in mouse embryos,
E11.5 [Reproducibility]
TAl1_1578
mm104
Enhancer (10.1±1.3)
Heart[7/7] | melanocytes[5/7] | liver[3/7]
TAL1_1496
hs796
Enhancer (6.3 ±0.9)
Forebrain [4/5]
TAL1_2105
mm291
Enhancer (6.3±1.3)
Heart [6/7]
TAL1_2302
hs1866
Enhancer (5.8±1.5)
Blood vessels [5/5]
TAL1_1123
hs1466
Enhancer (4.5±0.6)
Neural tube[8/8] | hindbrain (rhombencephalon)[8/8] |
midbrain (mesencephalon)[8/8] | dorsal root
ganglion[7/8] | forebrain[6/8] | limb[8/8] | branchial
arch[8/8] | heart[6/8]
TAL1_2750
hs1860
Enhancer (3.9±0.4)
Midbrain (mesencephalon) [5/5]
TAL1_3467
hs840
Enhancer (3.2±0.4)
Forebrain [10/10]
TAL1_1020
hs1385
Inactive (1.4±0.4)
Hindbrain (rhombencephalon)[3/5] | midbrain
(mesencephalon)[4/5]
TAL1_250
hs1862
Inactive (0.4±0.2)
Heart [6/6]
Figure S2: Enhancer activity of TAL1 OSs. (A) Illustration of expression vectors used in
transient transfection assays. In each expression vector, a TAL1 OS is inserted upstream of
a firefly luciferase reporter gene expressed from the human
A-globin
gene promoter
(HBG1pr). After transfection the expression level of the test construct is compared to that
from the parental vector, in both cases normalized to the expression of a co-transfection
control plasmid with the Renilla luciferase gene expressed from the promoter for a viral gene
encoding thymidine kinase (TKpr). (B) Comparison of the results of nine DNA segments
bound by TAL1 peaks tested in two enhancer assays (Tables S3-S4 in Additional file 2).
[0-0.1]kb
(0.1-1]
(1-2]
(2-5]
(5-50]
(50-500]
(500-2000]
(B)
C
Number of RefCRMs
5
0
10 15
Genome-wide distribution of
TAL1 OSs in each cluster (%)
0
20 40 60 80 100
(A)
1
2
3
7
4
5
6
Cluster no
16
13 12
5
12
5
2
1
1 2 3 4 5 6 7 8
Cluster no
8
(C)
(D)
(E)
100
80
60
40
20
50
40
Inactive Threshold Active
0
20
68
EP300 co-occupied TAL1OSs (%)
853
30
125 194
20
613
246
10
1347
40
60
80
343
608
214
374
817
197
324
505
0
221
GC-content of TAL1 OSs (%)
100
DHSs
EP300
Inactive Threshold Active
0
% of TAL1 OSs marked by DHSs, EP300
D
1
2
3
4
5
6
7
8
Cluster no
Figure S3: Classification of TAL1 OSs based on epigenetic features. (A) Distributions of
positions of TAL1 OSs in each cluster relative to the TSS of genes. (B) The numbers of
known reference cis-regulatory modules (CRMs) overlapping with TAL1 OSs in each cluster.
(C) Percentages of TAL1 OSs marked by DNaseI hypersensitive sites DHSs) and co-bound
by EP300; numbers of sites are over each column. (D,E) The percentage of GC-content (D)
and EP300 co-occupancy (E) is shown for the tested TAL1 OSs whose activities fall into
each of three activity categories.
1.0
1.5
2.0
0.1
TAL1 (RPM)
0.2
0.3
0.4
0.5
0
5
10
15
20
0
5
10
15
20
0
5
10
15
20
15
10
5
0.5
R=+0.04
R=-0.18
20
R=+0.1
R=-0.08
0
Fold change in activity
(A)
0.0
GATA1 (RPM)
0.2
0.4
0.6
0.8
0.2
H3K4me1 (RPM)
0.4
0.6
0.8
H3K4me3 (RPM)
Enh
Inac Thr
Inac Thr
Inac Thr
Enh
Enh
(B)
0.5
1.0
1.5
2.0
0.1
TAL1 (RPM)
0.3
0.4
0.5
“Good Binders-TAL1”
2.0
TAL1 signal at TAL1 peaks (RPM)
0
1
2
3
15
20
R=+0.32
0
5
10
15
10
5
1.5
-1
R=-0.11
0
1.0
-2
“Weak Binders-GATA1”
20
20
15
10
5
0.5
-3
log2(K4me1/K4me3) (RPM)
“Good Binders-GATA1”
R=-0.11
0.0
-4
GATA1 (RPM)
0
Fold change in activity
(C)
0.2
0.0
0.5
1.0
1.5
2.0
GATA1 signal at TAL1 peaks (RPM)
0.0
0.5
1.0
1.5
2.0
GATA1 signal at TAL1 peaks (RPM)
Figure S4: Signal strength of TAL1 binding, GATA1 binding, H3K4me1 and H3K4me3 enrichment
at TAL1 peaks (A) Relationship between enhancer activity and ChIP-seq signal strength values (RPM,
Reads Per Million) for TAL1, GATA1, H3K4me1 and H3K4me3 at 70 TAL1 OSs. (B) The distribution of
ChIP-seq signal strength of these features at 70 TAL1 OSs partitioned into three activity categories:
Inactive (Inac), Threshold (Thr) and Active Enhancers (Enh). (C) Relationship between enhancer activity
and ChIP-seq signal strength values of “good binders” (defined as those that pass the peak-calling
threshold) for TAL1 and GATA1 good binders versus that for GATA1 weak binders (signal is below the
threshold for peak-calling) at TAL1 peaks. Correlation coefficients (R) are given in each graph for (A) and
(C).
46
30
103
44
16
34
Enhancer
Threshold
Inactive
100
20
40
60
80
not overlap CAGE-tags
overlap CAGE-tags
0
Fraction of DNA segments
in each functional category (%)
B
Figure S5: Plotting fraction of DNA segments in each functional category that have or have
not overlap CAGE-tags.
DHS+ac+T
DHS+ac+T+P
DHS+ac+T+G
DHS+ac+T+G+P
DHS+m1+ac+T
DHS+m1+ac+T+P
DHS+m1+ac+T+G
DHS+m1+ac+T+G+P
DHS+m1+m3+ac-TF
DHS+m1+m3+ac+T
DHS+m1+m3+ac+T+P
DHS+m1+m3+ac+T+G
DHS+m1+m3+ac+T+G+P
DHS+m1+m3+T
DHS+m1+m3+T+P
DHS+m1+m3+T+G
DHS+m1+m3+T+G+P
DHS+m3+ac+T
DHS+m3+ac+T+P
DHS+m3+ac+T+G
DHS+m3+ac+T+G+P
DHS+m3+T
DHS+m3+T+P
DHS+m3+T+G+P
DHS+No HM+TF
DHSs
DHS+ac+P
DHS+ac+G
DHS+ac+G+P
DHS+m1+ac+P
DHS+m1+ac+G
DHS+m1+ac+G+P
DHS+m1+m3+P
DHS+m1+m3+G
DHS+m1+m3+G+P
DHS+m1+m3+ac
DHS+m1+m3+ac+P
DHS+m1+m3+ac+G
DHS+m1+m3+ac+G+P
DHS+m1+T
DHS+m1+T+P
DHS+m1+T+G
DHS+m1+T+G+P
DHS+m3+P
DHS+m3+G
DHS+m3+G+P
DHS+m3+ac-TF
DHS+m3+ac+P
DHS+m3+ac+G
DHS+m3+ac+G+P
DHS+T
DHS+T+P
DHS+T+G
DHS+T+G+P
DHS+G+P
DHS+m1+G+P
DHS+m3+ac
DHS+P
DHS+G
DHS+m1+P
DHS+m1+G
DHS+m1+m3-TF
DHS+m1+m3
DHS+m3-TF
DHS+CAGE tags
DHS+m3
DHS+m3+T+G
DHS+m1+ac-TF
DHS+m1+ac
DHS+Kac-TF
DHS+Kac
DHS+m1-TF
DHS+m1
DHS+No HM-TF
DHS+No HM-TF
DHS+T
DHS+m1
DHS+No HM+TF
DHS+ac
DHS+ac+G
DHS+m1+G
DHS+m1+ac
DHS+m1+ac+G
DHS+m1+T
DHS+m3
DHS+m3-TF
DHS+G
DHS+P
DHS+CAGE tags
DHS+ac-TF
DHS+ac+P
DHS+ac+G+P
DHS+ac+T
DHS+ac+T+P
DHS+ac+T+G
DHS+ac+T+G+P
DHS+m1-TF
DHS+m1+P
DHS+m1+G+P
DHS+m1+ac-TF
DHS+m1+ac+P
DHS+m1+ac+G+P
DHS+m1+ac+T
DHS+m1+ac+T+P
DHS+m1+ac+T+G
DHS+m1+ac+T+G+P
DHS+m1+m3
DHS+m1+m3-TF
DHS+m1+m3+P
DHS+m1+m3+G
DHS+m1+m3+G+P
DHS+m1+m3+ac
DHS+m1+m3+ac-TF
DHS+m1+m3+ac+P
DHS+m1+m3+ac+G
DHS+m1+m3+ac+G+P
DHS+m1+m3+ac+T
DHS+m1+m3+ac+T+P
DHS+m1+m3+ac+T+G
DHS+m1+m3+ac+T+G+P
DHS+m1+m3+T
DHS+m1+m3+T+P
DHS+m1+m3+T+G
DHS+m1+m3+T+G+P
DHS+m1+T+P
DHS+m1+T+G
DHS+m1+T+G+P
DHS+m3+P
DHS+m3+G
DHS+m3+G+P
DHS+m3+ac
DHS+m3+ac-TF
DHS+m3+ac+P
DHS+m3+ac+G
DHS+m3+ac+G+P
DHS+m3+ac+T
DHS+m3+ac+T+P
DHS+m3+ac+T+G
DHS+m3+ac+T+G+P
DHS+m3+T
DHS+m3+T+P
DHS+m3+T+G+P
DHS+T+G
DHSs
DHS+G+P
DHS+T+G+P
DHS+T+P
DHS+m3+T+G
Difference in Specificity
Difference in Sensitivity
Sensitivity
1.00
(A)
0.75
0.50
0.25
0.00
DHS
Regardless DHS
0.00
0.25
1-Specificity
0.50
0.75
1.00
0.10
0.05
(B)
0.00
-0.05
-0.10
0.4
(C)
0.2
0.0
Figure S6: Contribution of DHS to discriminatory power of epigenetic features. (A) General
comparison of discriminatory power of each feature and different combinations of them by a receiver-
operator characteristic (ROC) plot between two conditions: “Adding DHS to each feature or feature
combinations (illustrated by empty circles)” and “Regardless DHS (illustrated by circles in light blue
color”. (B,C) Difference between the two conditions (presence of DHS versus regardless DHS) in the
values for sensitivity (B) and specificity (C) of each feature or feature combinations.
3
2
1
0
-1
Expression in K562 cells
(A) A
Merged
GATA2
GATA1
TAL1
EP300
+ - + - + - + - + -
2
1
0
-1
Merged
GATA2
GATA1
TAL1
+ - + - + - + - + EP300
Merged
GATA2
GATA1
TAL1
EP300
+ - + - + - + - + -
-2
-2
-1
0
1
2
Expression in K562 cells
(scrambled)
3
(C)
C
-3
Expression in K562 cells
(B)B
Figure S7: Power of TF binding (+) versus no binding by TF (-) to identify enhancers.
High-throughput enhancer assays done in the studies by (A) Kheradpour et al. [27],
and (B, C) Kwasnieski et al. [28]. Distribution of expression levels of (A) DNA segments
centered on evolutionarily conserved GATA motif instances within enhancer chromatin states
[27], or (B, C) ENCODE DNA segments by histone modifications [28]. They are marked by
each TF co-occupancy individually and occupancy by merged TFs in K562 cells. The results
of scrambled DNA segments are shown in (C). (See Methods for details)
TAL1 OSs
(A)
Transient transfection
39 Active
Transgenic mice
43 Positive
24 Inactive
23 Negative
Discriminating Motif Enumerator (DME)
200 motifs enriched in actives
200 motifs enriched in positives
200 motifs enriched in inactives
200 motifs enriched in negatives
Select top 10 motifs in each set
MEME-TOMTOM
108 motifs matched to 63 protein binding sites from databases (E-value<1)
56 motifs in both enhancers and non-enhancers
41 motifs enriched only in enhancers
Enriched motifs in
Transient Transfection
Transgenic mice
Enhancer
Positive
Pos
Neg
30
20
10
0
0
50
150
250
350
0
Inactive
50
150
250
350
Negative
30
20
10
0
0
10
20
30
40
Inac
0
40
Enh
Frequency
10
20
30
20 40 60 80
0
G+C content (%)
40
(C)
40
(B)
11 motifs enriched only in non-enhancers
0
50
100
150
200
0
50
100
150
200
DME score
Figure S8: Identification of motifs enriched in the TAL1 bound enhancers and inactive
segments. (A) The overall procedure for identifying TF binding site motifs that contributes to
discrimination of TAL1 OSs by activity.
(B) Distribution of G+C content (%) in four data sets:
Enhancers and inactive regions identified in transient transfection assays, and positives and
negatives from transgenic mice assays. (C) Frequency of 200 motifs defined by DME2. The top 10
motifs (5% of 200 motifs) were used to run further analyses (Tables S10-S13 in Additional file 2).
(A)
(C)
(B)
Negative
17 (34.7%)
32 (65.3%)
5 (62.5%)
3 (37.5%)
Positive
SMAD1(+)
Expression in K562 cells
2
0
-2
SMAD1(-)
TAL1-GATA1
co-occupied segments
-)
+)
)(+ )
D1)(
D 1) (
AD 1
A
A
M
M
M
S
S
1or
1 &S )
)
L1& 3)
(TAL (144 (TA (130
( T A L (5 2
-4
Figure S9: Power of SMAD1 binding to TAL1 OSs to identify enhancers in human (A)
SMAD1 occupy the genome region with key regulators of the erythroid lineage (K562 cells).
Number of regions bound by SMAD1, TAL1 and GATA1 in K562 cells are shown. (B)
Partitions of 49 SMAD1-bound and eight SMAD1-unbound TAL1-GATA1 co-occupied
segments (K562 cells) by enhancer activity in transgenic mouse at E11.5day. (C) Distribution
of expression levels (high-throughput enhancer assay, [28]) of ENCODE DNA segments by
histone modification marked by TAL1-SMAD1 co-occupancy, TAL1 or SMAD1 occupancy,
and no occupancy by these two TFs in K562 cells.
(A)
(B)
Figure S10: The top gene ontology (GO) biological process (A) and Mouse Phenotype (B)
results of SMAD1 peaks in induced G1E-ER4 cells enriched single nearest gene within 1000
kb. The analyses were done using GREAT [62], Genomic Regions Enrichment Annotation
Tool. SMAD1 peaks (1,586) picked 1,262 (6%) of all genes.
(A)
Heart
Midbrain
Forebrain
Hindbrain
Blood vessels
Neural tube
Liver
Branchial arch
Limb
Melanocytes
Dorsal root
Facial mesenchyme
TAL1(+), SMAD1(+)
TAL1(+), SMAD1(-)
Somite
Other
0
(B)
5
10
15
Midbrain
Heart
Hindbrain
Neural tube
Forebrain
Limb
Branchial arch
Dorsal root ganglion
Somite
Blood vessels
Trigeminal V
Nose
Eye
Cranial nerve
Tail
Liver
Facial mesenchyme
Ear
Genital tubercle
Pancreas
Melanocytes
Other
0
20
25
30
TAL1(+), SMAD1(+)
TAL1(+), SMAD1(-)
10
20
30
40
50
Figure S11: Distribution of tissues depicting enhancement by the TAL1 OSs in the presence
(+) or absence (-) of SMAD1 occupancy in (A) G1E-ER4+E2 cells (mouse) and (B) K562
cells (human).
Figure S12: TF occupancy in human hematopoietic cells within the Gypc gene, which encodes
the erythroid membrane protein glycophorin C. ChIP-seq density plots were obtained from CODEX
compendium as bigWig format [63], and displayed on the UCSC Genome Browser. Binding profiles
depicted are for TAL1 (GEO sample numbers: GSM614004, GSM1278241, GSM651545), GATA1
(GSM467647, GSM1278240, GSM651547, GSM610335),
GATA2 (GSM467648, GSM1097883),
SMAD1 (GSM722397, GSM722397), STAT4 (GSM945607, GSM945608, GSM945610_11), IRF2
(GSM970261), and IRF5 (GSM945603_04, GSM945605), in descending order. A strong peak for
SMAD1 overlapping a TAL1 peak is outlined. No peaks were detected in this locus for STAT2, IRF2, or
IRF5, but overlap with TAL1 was observed in other locations.
(A)
(B)
(C)
(D)
(E)
(F)
Figure S13: Effect of binding by IRF2, STAT1, FOXO1 or SMAD3 to TAL1 OSs to identify
enhancers in hematopoietic cells. IRF2 in pro-erythroblasts occupy TAL1 binding sites in
proerythroblats (A) and K562 cells (B). STAT1 in Macrophages (C), STAT1 in T-cells (D), FOXO1 in
CD4+ cells (E), and SMAD3 in B cells (F), occupy the genome region with TAL1 in G1E-ER4 cells.
Number of regions bound by TFs are shown. (A-B in human cells; C-F in mouse cells).
60
0
20
40
60
40
20
0
0.4
0.6
0.8
1.0
0.0
0.2
0.4
0.6
0.8
1.0
0.0 0.5 1.0 1.5 2.0 2.5
40
20
0
20
40
60
0.2
60
0.0
0
Frequency of TAL1 OSs in
Transgenic mice
Transient transfection
(A)
0.0 0.5 1.0 1.5 2.0 2.5
0.0
0.5
1.0
1.5
PhyloP score
PhyloP score
PhastCons score
(B)
0.0 0.2 0.4 0.6 0.8 1.0
PhastCons score
-2 -1
0
1
2
3
4
-2 -1
0
1
2
3
4
Fold change in activity (log2)
Figure S14: Contribution of conservation to enhancer activity of TAL1 OSs. (A)
Frequency of TAL1 OSs tested in the two enhancer assays based on PhastCons (left) and
PhyloP (right) scores (B) Correlation of TAL1 OSs between the conservation scores
(PhastCons and PhyloP ) and enhancer activity in transient transfection assay.
(A)
(B)
Figure S15: Contribution of WGATAR motif preservation to enhancer activity of TAL1
OSs. (A) Examples of occurrence of a deeply preserved ad a lineage-specific WGATAR
motif in a TAL1 peak, discovered by CladiMo. (B) The distribution of enhancer activities for
DNA segments is represented as dot plots and partitioned by preservation on the GATA1
binding site motifs. The total numbers of occupied segments bound by GATA1 (left) or cobound by GATA1 and TAL1 (right) in each category are given at the bottom of plots. The red
internal line indicates the median of enhancer activity in each category (no motif, non
preserved motif, or preserved motif between human and mouse).