Using an ensemble of statistical metrics to quantify large sets of plant transcription factor binding sites

Hosseini, Parsa; Ovcharenko, Ivan; Matthews, Benjamin F

doi:10.1186/1746-4811-9-12

Plant Methods

Table 4 Various metrics infer differing magnitudes of TFBS over-representation

From: Using an ensemble of statistical metrics to quantify large sets of plant transcription factor binding sites

	Metrics								TFBS raw-abundance
TF	LP	CO	JAC	LI	CF	K	PHI	p-value	Suppressed	Induced
ABF1	20	39	39	20	20	3	2	8.211e-274	130	169
ABFS	9	9	10	9	9	16	12	2.385e-31	10	20
ABI3/FUS3	67	19	17	67	67	41	58	3.036e-47	14	7
ABI4(2)	64	34	33	64	64	64	67	4.465e-172	66	43
AG	14	20	21	14	14	13	18	4.611e-82	30	42
AGP1	48	57	56	48	48	58	49	2.412e-720	427	398
ALFIN1	34	58	57	34	34	34	34	1.580e-731	440	426
ARF1	65	29	24	65	65	57	62	1.243e-113	40	25
ARR10	39	65	65	39	39	43	39	1.836e-895	579	552
ARR2	69	27	22	69	69	60	69	4.028e-99	33	15
ATHB-5	43	68	68	43	43	49	43	1.542e-901	584	555
ATHB1	40	67	67	40	40	45	40	3.162e-901	584	556
ATHB5-1	63	21	20	63	63	44	55	3.202e-78	26	18
ATHB5-2	37	60	60	37	37	37	37	9.771e-769	470	452
ATHB6	27	23	25	27	27	29	32	3.067e-109	41	46
ATHB9	53	38	36	53	53	55	52	2.105e-225	95	81
AtLEC2	55	51	51	55	55	68	61	1.066e-611	336	284
ATML1/PDF2	71	18	11	71	71	54	71	8.730e-38	10	1
AtMYB2	29	33	34	29	29	23	31	1.606e-170	70	76
AtMYB77	60	32	31	60	60	56	57	2.955e-141	53	40
AtMYC2	2	2	2	2	2	30	8	0.0002735	1	7
AtSPL3	30	45	46	30	30	8	26	7.997e-426	220	236
BLR/RPL/PNY	35	61	61	35	35	35	35	1.444e-777	478	462
bZIP910(2)	10	12	16	10	10	14	11	6.060e-42	14	26
bZIP911	12	11	13	12	12	19	14	4.350e-37	12	21
bZIP911(1)	11	10	12	11	11	20	13	2.529e-34	11	20
bZIP911(2)	18	13	14	16	16	32	29	3.730e-38	12	16
CBF	43	68	68	43	43	49	43	1.542e-901	584	555
CDC5	4	4	4	4	4	18	3	1.343e-10	3	13
DOF2	42	71	71	42	42	48	42	1.259e-902	585	556
DPBF1/2	51	55	55	51	51	66	54	1.857e-712	418	379
E2Fa	70	13	9	70	70	38	64	8.059e-24	6	1
E2Fc/d	1	1	1	1	1	26	5	0.0003077	1	8
EmBP-1	25	43	43	25	25	5	17	3.316e-397	203	228
GAMYB	47	59	59	47	47	53	47	2.040e-743	447	422
Gamyb	58	28	26	58	58	40	50	6.185e-120	44	36
GATA-1	17	24	28	18	18	12	16	5.923e-120	47	62
GATA-1/2/3/4	16	15	18	17	17	28	27	9.291e-54	18	24
GT-3b	13	25	29	13	13	7	7	1.244e-128	52	76
HAHB4	46	64	64	46	46	52	46	1.038e-891	575	546
HAT5	43	68	68	43	43	49	43	1.542e-901	584	555
HSE	19	26	30	19	19	11	15	1.641e-130	52	68
HVH21	41	66	66	41	41	46	41	3.881e-900	583	555
HY5	6	8	8	6	6	21	10	6.247e-20	6	15
ID1	28	31	32	28	28	27	33	4.015e-146	58	63
MYB.PH3(1)	56	41	41	56	56	61	56	4.159e-333	154	130
MYB.PH3(2)	52	49	49	52	52	62	53	6.937e-564	306	276
MYB98	62	36	35	62	62	65	65	2.648e-210	85	60
O2	33	56	58	33	33	6	28	2.967e-731	446	457
OsbHLH66	26	40	40	26	26	9	20	1.723e-308	147	165
OsCBT	3	3	3	3	3	24	6	1.543e-7	2	10
P	57	52	52	57	57	71	66	2.571e-629	347	286
PCF2	61	47	44	61	61	70	70	3.566e-441	215	160
PCF5	59	48	48	59	59	69	68	2.612e-498	254	201
PEND	31	35	37	31	31	15	30	3.825e-230	101	108
PIF3(2)	21	22	23	21	21	17	25	4.178e-99	37	46
RAP2.2	66	30	27	66	66	59	63	1.614e-125	45	28
RAV1(1)	49	54	53	49	49	63	51	1.957e-688	400	366
RAV1(2)	38	62	62	38	38	39	38	1.073e-854	543	519
STF1	24	37	38	24	24	10	22	1.243e-242	109	124
TAC1	68	17	15	68	68	42	59	1.479e-44	13	6
TaMYB80	54	50	50	54	54	67	60	2.700e-594	324	276
TBP	36	63	63	36	36	36	36	1.545e-881	568	547
TEIL	50	42	42	50	50	47	48	8.458e-340	160	146
TGA1	23	46	47	23	23	2	9	3.416e-468	253	293
TGA1a	32	53	54	32	32	4	23	3.325e-688	413	433
WRKY11	7	7	7	8	8	31	19	2.346e-14	4	9
WRKY18/40/62	7	6	6	7	7	33	21	2.879e-11	3	7
WRKY26/38/43	15	16	19	15	15	25	24	4.164e-56	19	26
WRKY6	5	5	5	5	5	22	4	1.091e-10	3	12
ZAP1	22	44	45	22	22	1	1	2.468e-415	219	268

Promoter sequences from the top 600 induced and top 600 suppressed genes 10 dai were identified and their TFBS abundances quantified using Marina. A catalog of pre-assembled DNA motifs (1,160 motifs) and PWMs (80 matrices) accompanied such groups.
A total of 71 over-represented TFBSs were identified. Of these N TFBSs, magnitude of over-representation is ranked from 1 to N such that the most over-represented are close to 1 while the least over-represented are close to N. Since TFBS models can vary across source-organisms, certain over-represented TFBSs were found multiple times (i.e. GAMYB, bZIP911, and ATHB5). Furthermore, not all metrics rank the same. As a result, manually deducing degree of TFBS over-representation can be a challenging task. IPF-standardization is designed to circumvent such a scenario.

Back to article page

ISSN: 1746-4811

Contact us

Submission enquiries: journalsubmissions@springernature.com