Global Methylation Profiling for Risk Prediction of Prostate Cancer

Saswati Mahapatra; Eric W. Klee; Charles Y.F. Young; Zhifu Sun; Rafael E. Jimenez; George G. Klee; Donald J. Tindall; Krishna Vanaja Donkena

doi:10.1158/1078-0432.CCR-11-2090

DOI: 10.1158/1078-0432.CCR-11-2090 Published May 2012

Article Figures & Data

Figures

Tables
Additional Files

Download figure
Open in new tab
Download powerpoint
Figure 1.
Methylation of genes in prostate cancer tissues, analyzed using Infinium Methylation27 BeadChips. Cluster diagram depicting genes that distinguish 2 contrast groups. A, genes that are significantly methylated in prostate cancer tissues (n = 158) compared with normal prostate (n = 34); normal (N), nonrecurrence (NR), biochemical recurrence (BR), local recurrence (LR), and systemic recurrence (SR) prostate cancer tissues. B, hypermethylated genes in tissues of patients with recurrent prostate cancer (n = 98) compared with nonrecurrent patients (n = 60). C, significance of genes methylated in prostate cancer tissues of patients with clinical recurrence (n = 59) compared with patients with only biochemical recurrence (n = 39). D, genes that are methylated in tissues of patients with systemic recurrence (n = 23) compared with local recurrence (n = 36). Each row represents a gene and each column a tissue sample. Methylation is color scaled from white to black such that white represents zero (no methylation detected) and black represents more than 99% (input DNA is methylated), respectively, relatively to the median of the reference pool. Color saturation is proportional to the magnitude of the difference from the mean. Each gene is labeled by its gene name.
Download figure
Open in new tab
Download powerpoint
Figure 2.
Representative pyrogram of genes methylated in prostate cancer tissues. The sequence in the top part of each pyrogram represents the sequence under investigation. The sequence below the pyrogram indicates the sequentially added nucleotides. The gray regions highlight the analyzed C/T sites, with percentage values for the respective cytosine above them. A, pyrogram of genes HIF3A, HAAO, RARβ, and GSTP1 methylated in prostate cancer tissues. B, pyrogram of genes RUNX3, HS3ST2, CRIP1, and FLNC methylated in prostate cancer tissues of patients who had recurrence.

Table 1.

Prostate cancer tissues used for methylation microarray analysis

	Nonrecurrent	Biochemical recurrence	Local recurrence	Systemic recurrence
Number of patients	75	43	44	36
Age, y	62.61 ± 8.04	64.04 ± 5.95	63.31 ± 6.03	64.41 ± 7.57
Preoperative PSA, ng/mL
<4	17 (22.6)	1 (2.3)	2 (4.5)	3 (8.3)
4–10	45 (60)	22 (51.1)	19 (43.1)	13 (36.1)
>10	12 (16)	17 (39.5)	15 (34.09)	15 (13.8)
n.a.	19 (1.3)	3 (6.9)	8 (18.1)	5 (13.8)
Gleason score
6	23 (30.6)	4 (9.3)	3 (6.8)	0
7	34 (45.3)	19 (44.1)	27 (61.3)	11 (30.5)
8	8 (10.6)	8 (18.6)	6 (13.6)	5 (13.8)
9	10 (13.3)	12 (27.9)	8 (18.1)	15 (41.6)
10	0	0	0	5 (13.8)
TNM
T2a	19 (25.3) [4]	7 (16.2) [1]	11 (25) [3]	1 (2.7)
T2b	30 (40) [16]	14 (32.5) [3]	16 (36.3) [9]	6 (16.6) [4]
T3a	20 (26.6) [11]	10 (23.2) [7]	8 918.1) [4]	12 (33.3) [5]
T3b	6 (8) [3]	12 (27.9) [9]	7 (15.9) [3]	11 (30)[7]
T4	0	0	2 (4.5) [1]	6 (16.6) [4]
Ploidy
Diploid	38 (50.6)	19 (44.1)	17 (38.6)	15 (41.6)
Tetraploid	29 (38.6)	19 (44.1)	22 (50)	13 (36.1)
Aneuploid	8 (10.6)	5 (11.6)	5 (11.3)	8 (22.2)
GPSM score
<10	47 (62.2)	15 (34.8)	10 (22.7)	6 (16.6)
10–12	22 (29.3)	14 (32.5)	20 (45.4)	14 (38.8)
>12	5 (6.6)	11 (25.5)	6 (13.6)	12 (33.3)
n.a.	1 (1.3)	3 (6.9)	8 (18.2)	5 (13.8)
PSA progression, y	0	2.33 ±1.77	0.77 ± 0.42	1.39 ± 1.54
Local progression, y	0	0	2.31 ± 1.18	3.27 ± 2.99
Systemic progression, y	0	0	0	2.59 ±3.41
Follow-up, y	6.25 ± 1.48	5.87 ± 1.41	4.31 ± 1.69	4.43 ± 3.97
Prostate death	0	0	0	6
Other death	0	6	2	7

NOTE: Mean ± S.D are shown for age, years for PSA, local, and systemic progression and follow-up years of patients with prostate cancer. Value in parentheses indicates the percentage of cases representing the group. Value in square brackets indicates the node-positive cases. Matched normal tissues from 40 patients [age (mean ± SD) = 63.15 ± 7.28] were used.

Abbreviation: n.a., not available.

Table 2.

Significance of differentially methylated genes in prostate cancer tissues, analyzed using Infinium Methylation27 BeadChips

Gene	Sensitivity (95% CI)	(95% CI)	AUC	t Test (P)	Transcription start site	Strand	CpG island location
A, Genes that are significantly methylated in prostate cancer tissues (n = 158) compared with normal prostate (n = 34)^a
A0X1	96.2 (91.9–98.5)	100 (89.7–100)	0.992	<0.0001	41	+	2: 201,158,654–201,159,798
CYBA	95.5 (91.1–98.2)	97.1 (84.7–99.9)	0.988	<0.0001	83	−	16: 87,244,422–87,245,165
EDG3	93.0 (87.9–96.4)	97.1 (84.7 to 99.9)	0.97	<0.0001	363	+	9: 90,794,956–90,796,827
ELF4	88.6 (82.6–93.1)	100 (89.7–100)	0.994	<0.0001	600	−	X: 129,071,186–129,073,330
EPB41L3	93.7 (88.7–96.9)	100 (89.7–100)	0.978	<0.0001	185	−	18: 5,532,741–5,534,303
FLJ12056	94.3 (89.5–97.4)	100 (89.7–100)	0.978	<0.0001	316	+	2: 71,058,922–71,060,108
FLJ90650	96.2 (91.9–98.6)	100 (89.7–100)	0.994	<0.0001	449	+	5: 115,324,903–115,327,577
FLT4	99.4 (96.5–99.9)	100 (89.7–100)	0.999	<0.0001	691	−	5: 180,008,152–180,010,059
GAS6	96.2 (91.9–98.6)	100 (89.7–100)	0.99	<0.0001	964	+	13: 113,546,578–113,548,278
GRASP	96.2 (91.9–98.6)	100 (89.7–100)	0.993	<0.0001	159	+	12: 50,686,496–50,688,060
GSTP1	91.8 (86.3–95.6)	100 (89.7–100)	0.995	<0.0001	755	+	11: 67,106,980–67,108,634
HAAO	93.0 (87.9–96.5)	100 (89.7–100)	0.979	<0.0001	454	−	2: 42,873,041–42,874,496
HIF3A	95.6 (91.1–98.2)	100 (89.7–100)	0.989	<0.0001	153	+	19: 51,491,754–51,492,473
HOXC11	89.2 (83.3–93.6)	97.1 (84.7–99.9)	0.985	<0.0001	20	+	12: 52,652,893–52,656,105
LEP	98.1 (94.6–99.6)	97.1 (84.7–99.9)	0.997	<0.0001	51	+	7: 127,667,928–127,668,724
MGC39606	95.6 (91.1–98.2)	100 (89.7–100)	0.992	<0.0001	634	+	X: 134,383,127–134,384,321
MOBKL2B	92.4 (87.1–96.0)	100 (89.7–100)	0.97	<0.0001	851	−	9: 27,517,915–27,520,017
RAB34	91.8 (86.3–95.6)	94.1 (80.3–99.3)	0.976	<0.0001	157	−	17: 24,068,248–24,069,280
RARβ	93.0 (87.9–96.5)	100 (89.7–100)	0.986	<0.0001	352	+	3: 25,444,208–25,445,101
RHCG	95.6 (91.1–98.2)	100 (89.7–100)	0.993	<0.0001	110	−	15: 87,840,274–87,841,084
RND2	93.0 (87.9–96.5)	100 (89.7–100)	0.996	<0.0001	180	+	17: 38,429,704–38,431,262
SLC34A2	91.1 (85.6–95.1)	94.1 (80.3–99.3)	0.966	<0.0001	44	+	4: 25,266,077–25,266,795
SPATA6	95.6 (91.1–98.2)	100 (89.7–100)	0.991	<0.0001	595	−	1: 48,709,694–48,710,842
TPM4	94.9 (90.3–97.8)	100 (89.7–100)	0.977	<0.0001	306	+	19: 16,047,744–16,049,758
ZNF154	97.5 (93.6–99.3)	94.1 (80.3–99.3)	0.99	<0.0001	100	−	19: 62,911,404–62,912,681
B, Hypermethylated genes in tissues of patients with recurrent prostate cancer (n = 98) compared with nonrecurrent cancer (n = 60)^b
ACTL6B	65.3 (55.1–74.6)	66.7 (53.3–78.3)	0.702	<0.0001	136	−	7: 100,091,590–100,092,362
AEBP1	63.3 (52.9–72.8)	60.0 (46.5–72.4)	0.646	<0.0021	38	+	7: 44,110,152–44,111,361
AMID	77.5 (68.0–85.4)	76.7 (64.0–86.6)	0.843	<0.0001	25	−	10: 71,562,005–71,562,897
CD8A	75.5 (65.8–83.6)	71.7 (58.6–82.5)	0.829	<0.0001	114	−	2: 86,869,360–86,871,837
CRIP1	61.2 (50.8–70.9)	60.0 (46.5–72.4)	0.638	<0.0037	93	+	14: 105,023,580–105,025,363
FLJ30934	80.6 (71.4–87.9)	78.3 (65.8–87.9)	0.862	<0.0001	82	+	11: 65,357,151–65,358,255
FLNC	79.6 (70.3–87.1)	81.7 (69.6–90.5)	0.864	<0.0001	131	+	7: 128,257,322–128,258,463
FMOD	76.5 (66.9–84.5)	68.3 (55.0–79.7)	0.748	<0.0001	97	−	1: 201,586,895–201,587,112
FOXE3	65.3 (55.0–74.6)	63.3 (49.9–75.4)	0.697	<0.0001	570	+	1: 47,654,351–47,655,785
GAS7	66.3 (56.1–75.6)	66.7 (53.3–78.3)	0.744	<0.0001	395	−	17: 10,041,659–10,043,517
GDPD5	70.4 (60.3–79.2)	65.4 (51.6–76.8)	0.713	<0.0001	408	−	11: 74,913,525–74,915,608
HS3ST2	65.3 (55.0–74.6)	75.0 (62.1–85.3)	0.798	<0.0001	257	+	16: 22,732,005–22,734,135
LOC349136	67.3 (57.1–76.5)	65.0 (51.6–76.9)	0.742	<0.0001	479	−	7: 150,736,827–150,739,238
NEUROG1	66.3 (56.1–75.6)	70.0 (56.8–81.2)	0.731	<0.0001	19	−	5: 134,898,284–134,900,130
PLTP	68.4 (58.2–77.4)	75.1 (62.1–85.3)	0.709	<0.0001	165	−	20: 43,972,342–43,974,350
PTGER2	66.3 (56.1–75.6)	66.7 (53.3–78.3)	0.739	<0.0001	199	+	14: 51,850,265–51,852,038
RASGRF2	69.4 (59.3–78.3)	71.7 (58.6–82.5)	0.763	<0.0001	106	+	5: 80,291,476–80,292,927
RUNX3	67.4 (57.1–76.5)	71.7 (58.6–82.6)	0.746	<0.0001	528	−	1: 25,127,692–25,131,906
SIX6	65.3 (55.0–74.6)	68.3 (55.0–79.7)	0.745	<0.0001	185	+	14: 60,045,112–60,046,647
SLC9A3	69.4 (59.3–78.3)	70.0 (56.8–81.2)	0.713	<0.0001	178	−	5: 576,269–577,990
SPSB4	62.2 (51.9–71.8)	61.7 (48.2–73.9)	0.658	<0.0008	91	+	3: 142,252,136–142,254,676
SRD5A2	70.4 (60.3–79.2)	71.7 (58.6–82.5)	0.748	<0.0001	125	−	2: 31,658,633–31,660,643
SUSD3	64.3 (54.0–73.7)	63.3 (49.9–75.4)	0.652	<0.0014	560	+	9: 94,860,370–94,861,840
SYT10	66.3 (56.1–75.6)	65.4 (51.6–76.8)	0.714	<0.0001	254	−	12: 33,482,438–33,484,467
TMEM74	66.3 (56.1–75.6)	68.3 (55.0–79.7)	0.664	<0.0005	14	−	8: 109,868,266–109,869,213
C, Significance of genes methylated in prostate cancer tissues of patients with clinical recurrence (n = 59) compared with patients with only biochemical recurrence (n = 39)^c
CHST7	69.5 (56.1–80.8)	64.1 (47.2–78.8)	0.668	<0.0050	255	+	X: 46,317,905–46,319,911
LMX1B	71.2 (57.9–82.2)	64.1 (47.2–78.8)	0.684	<0.0021	106	+	9: 128,412,539–128,418,074
PHLDA 3	61.0 (47.4–73.4)	59.0 (42.1–74.4)	0.606	<0.0772	348	−	1: 199,704,096–199,705,294
RAFTLIN	62.7 (49.1–75.0)	64.1 (47.2–78.8)	0.668	<0.0050	89	−	3: 16,529,156–16,530,542
RASGRF2	66.1 (52.6–77.9)	69.2 (52.4–82.9)	0.690	<0.0015	106	+	5: 80,291,476–80,292,927
TNFRSF10D	62.7 (49.2–74.9)	66.6 (49.7–80.9)	0.700	<0.0008	25	−	8: 23,076,422–23,077,641
ZNF135	59.3 (45.7–71.9)	61.5 (44.6–76.6)	0.664	<0.0061	144	+	19: 63,262,091–63,264,376
D, Genes that are methylated in tissues of patients with systemic recurrence (n = 23) compared with local recurrence (n = 36)^d
ALPL	65.2 (42.7–83.6)	63.8 (46.2–79.1)	0.683	<0.0182	269	+	1: 21,707,968–21,709,033
AMPH	65.2 (42.7–83.6)	75.0 (57.8–87.8)	0.77	<0.0005	35	−	7: 38,636,784–38,637,861
BCDIN3	78.2 (56.3–92.5)	83.3 (67.1–93.6)	0.861	<0.0001	873	+	7: 99,863,663–99,866,414
BCL11B	82.6 (61.2–95.0)	80.5 (63.9–91.8)	0.797	<0.0001	274	−	14: 98,806,507–98,807,587
BRD4	73.9 (51.6–89.7)	69.4 (51.8–83.6)	0.724	<0.0038	665	−	19: 15,252,811–15,253,031
C18orf34	78.2 (56.3–92.5)	77.7 (60.8–89.8)	0.806	<0.0001	174	−	18: 29,274,320–29,275,128
DCAMKL1	82.6 (61.2–95.5)	91.6 (77.5–98.2)	0.874	<0.0001	734	−	13: 35,602,437–35,603,802
FGF5	91.3 (71.9–98.9)	83.3 (67.1–93.6)	0.911	<0.0001	544	+	4: 81,405,716–81,407,598
FLJ42486	82.6 (61.2–95.0)	77.7 (60.8–89.8)	0.834	<0.0001	99	−	14: 104,141,516–104,142,673
JAM2	86.9 (66.4–97.2)	86.1 (70.5–95.3)	0.936	<0.0001	587	+	21: 25,933,195–25,934,679
LHX9	82.6 (61.2–95.0)	69.4 (51.8–83.6)	0.822	<0.0001	106	+	1: 196,146,152–196,148,467
LOC283537	78.2 (56.3–92.5)	75.0 (57.8–87.8)	0.811	<0.0001	363	−	13: 28,190,499–28,191,650
LRAT	65.2 (42.7–83.6)	72.2 (54.8–85.8)	0.727	<0.0034	462	+	4: 155,882,485–155,885,476
PDE4B	91.3 (71.9–98.9)	77.7 (60.8–89.8)	0.856	<0.0001	423	+	1: 66,030,397–66,031,906
POU3F3	91.3 (71.9–98.9)	75.2 (57.8–87.8)	0.865	<0.0001	248	+	2: 104,835,039–104,840,521
PTGS2	60.8 (38.5–80.2)	80.5 (63.9–91.8)	0.688	<0.0153	37	−	1: 184,915,751–184,916,808
RASGRF2	73.9 (51.6–89.7)	75.4 (57.8–87.8)	0.788	<0.0002	106	+	5: 80,291,476–80,292,927
SLC27A6	78.2 (56.3–92.5)	69.4 (51.8–83.6)	0.742	<0.0017	22	+	5: 128,328,505–128,329,501
SLC03A1	73.9 (51.6–89.7)	83.3 (67.1–93.3)	0.797	<0.0001	333	+	15: 90,196,981–90,198,831
SPSB4	82.6 (61.2–95.0)	80.6 (63.9–91.8)	0.824	<0.0001	91	+	3: 142,252,136–142,254,676
STAT3	69.5 (47.0–86.7)	69.4 (51.8–83.6)	0.744	<0.0017	757	−	17: 37,794,547–37,794,828
SYN2	73.9 (51.6–89.7)	80.5 (63.9–91.8)	0.809	<0.0001	406	+	3: 12,020,118–12,021,768
TACR3	69.5 (47.0–86.7)	72.2 (54.8–85.8)	0.704	<0.0086	346	−	4: 104,860,449–104,860,859
TIRAP	73.9 (51.6–89.7)	69.4 (51.9–83.6)	0.779	<0.0003	645	+	11: 125,657,511–125,658,777
WNT11	65.22 (42.7–83.6)	63.9 (46.2–79.2)	0.753	<0.0011	221	−	11: 75,594,610–75,600,053

NOTE: Performance of the genes was analyzed by AUC of the genes for the 2 contrast groups. Sensitivity and specificity are reported for the threshold that minimized sensitivity–specificity. The P value indicates the statistical significance. Location of the CpG site used for analysis is given with respect to the distance from the transcription start site. The CpG island location on the respective chromosome is shown.

Abbreviation: CI, confidence interval.

↵^at test P value is obtained by comparing tumor (n = 158) and benign tissues (n = 34).
↵^bt test P value is obtained by comparing prostate cancer tissues of patients with recurrence (n = 98) and without recurrence (n = 60).
↵^ct test P value is obtained by comparing prostate cancer tissues of patients with clinical recurrence (n = 59) and biochemical recurrence (n = 39).
↵^dt test P value is obtained by comparing prostate cancer tissues of patients with systemic recurrence (n = 23) and local recurrence (n = 36).

Table 3.

Forward and reverse primers used for PCR amplification and the sequencing primer used for pyrosequencing reactions

Gene	Annealing temp, °C	Forward primer	Reverse primer	Sequencing primer
HIF3A	56.1	GGTAGGAGTTTTGGGAATTGG	*-CCACCCCTACAATCCCTAA	GGGTGAGAGTTAAGA
HAAO	56.1	AGGTTTGGGGTTTATTTTGAAT	*-TAAAAATCCCAACACCTAAAAACAC	TTTTAGATGGGAAAGTTAAATT
RARβ	56.1	GTGTTTTTAAAATGAGTAGGGGAGGA	CCATTCTCTATTCTACAATTTAAAACTTAC	TTATTTGAGGAGGTTTAGTTTGGAA
GSTP1	57.5	*-AGGGAGTAAATAGATAGTAGGAAGAG	CCTCTCCCCTACCCTATAAA	AATCCCAACCATCTTAAAAAATTTC
CRIP1	59.2	TGTTTTAAGTGTAATAAGGAGGTGTAT	*-ACAAAAACCTCCACCCTC	AATAAGGAGGTGTATT
RUNX3	57.5	Hs_RUNX3_01_PM00000112
HS3ST2	59.2	GATTATAGGGGAGGGA11111GGAGAA	*-ACACTTACCTATATTACAACTTCCTATAC	GGGA11111GGAGAAG
FLNC	59.2	Hs_FLNC_03_PM00127330
RASGRF2	56.1	*-AAGTAGAAGAGTATATTTTGGTAGAG	AACACCCCCTATACCTAACCTTTCTAAC	AAACCCCACTTCCTAAATAAAAAAA
PHLDA3	57.5	GGGGTTTTATGGAGTGTTAAGTAATTAGT	*-AACAACCCAAACCTTCTAACT	TTTGAAGTTTAATATGTAATTGGTG
TNFRSF10D	59.2	*-ATGGTGGTTAGGTTGGTTTTTAA	CACTAAACTACCCCAAAAAAAATTCC	AACCTCAAAAAACCAC
ZNF135	57.5	GGI I I I I 1 AG 11 GAG 1A1 AG 1 GAG IGGG	*-CCTAAAAAAACCCCTTTTTCTCC	GTATAGTGAGTGGGG
BCL11B	57.5	ATATTAATTAGGGTAGAGGAGGAGGT	*-ATAATAAACTCCCTCTAAAACAAATACT	AGATGTAGGGGTAATGTTT
POU3F3	59.2	*-GGGGTAGGGTTGTGATTTATTG	AAAACCCCTCCTCTATCAAAATAATCCAAA	CCAAAACCACTTCATACAAAAAAC

NOTE: Annealing temperature used for PCR amplification is shown. RUNX3 and FLNC PyroMark assays were obtained from Qiagen.

*Indicates the PCR primer with biotinylation.

Table 4.

Validation of methylation by pyrosequencing in an independent batch tissue samples from 20 patients representing each group

Gene	Sensitivity (95% CI)	Specificity (95% CI)	AUC	t test (P)
Group 1^a
HIF3A	90.0 (68.3–98.77)	100 (83.1–100)	0.932	0.0001
HAAO	89.4 (66.8–98.7)	100 (83.1–100)	0.947	0.0001
RARβ	84.2 (60.4–96.6)	100 (82.3–100)	0.912	0.0001
GSTP1	94.7 (73.9–99.8)	100 (82.3–100)	0.977	0.0001
Group 2^b
CRIP1	65.0 (40.7–84.6)	65.6 (40.7–84.6)	0.727	0.0139
RUNX3	70.4 (45.7–88.1)	75.3 (50.9–91.3)	0.788	0.0018
HS3ST2	65.0 (40.8–84.6)	60.0 (36.0–80.9)	0.733	0.0115
FLNC	70.3 (45.7–88.1)	60.4 (36.0–80.8)	0.660	0.0835
RASGRF2	75.7 (50.9–91.3)	55.2 (31.5–76.9)	0.682	0.0515
Group 3^c
PHLDA3	65.6 (40.7–84.6)	65.0 (40.7–84.6)	0.73	0.0129
RASGRF2	75.4 (50.9–91.3)	60.3 (36.0–80.8)	0.761	0.0047
TNFRSF10D	60.8 (36.0–80.8)	75.6 (50.9–91.3)	0.692	0.0373
ZNF135	50.0 (27.2–72.8)	50.0 (27.2–72.8)	0.555	0.5518
Group 4^d
BCL11B	75.2 (50.9–91.3)	60.0 (36.0–80.8)	0.741	0.0091
POU3F3	65.5 (40.7–84.6)	70.7 (45.7–88.1)	0.701	0.0295
RASGRF2	70.6 (45.7–88.1)	75.4 (50.9–91.3)	0.748	0.0071

NOTE: Performance of the genes was analyzed by AUC of the genes for the 2 contrast groups. Sensitivity and specificity are reported for the threshold that minimized sensitivity–specificity. The P value indicates the statistical significance. t test p value is obtained by comparing 20 tissues in each group.

Abbreviation: CI, confidence interval.

↵^aGenes that are significantly methylated in prostate cancer tissues compared with normal prostate.
↵^bHypermethylated genes in tissues of patients with recurrent prostate cancer compared with nonrecurrent tissues.
↵^cSignificance of genes methylated in prostate cancer tissues of patients with clinical recurrence compared with patients with only biochemical recurrence.
↵^dGenes that are methylated in tissues of patients with systemic recurrence compared with local recurrence.

Additional Files

Figures
Tables

Supplementary Data

Files in this Data Supplement:

Supplementary Table 1A - PDF file - 419K, Significantly methylated genes in prostate cancer tissues from the binary group comparisons
Supplementary Table 6 - PDF file - 70K, Significance of PTGS2 methylation in different groups of prostate cancer tissues
Supplementary Table 1B - PDF file - 270K, Significantly methylated genes in prostate cancer tissues from the binary group comparisons
Supplementary Table 1C - PDF file - 112K, Significantly methylated genes in prostate cancer tissues from the binary group comparisons
Supplementary Table 1D - PDF file - 240K, Significantly methylated genes in prostate cancer tissues from the binary group comparisons
Supplementary Table 4A - PDF file - 269K, Significance of methylation in prostate cancer tissue samples from 4 microarrays with the batch effect
Supplementary Table 4B - PDF file - 273K, Significance of methylation in prostate cancer tissue samples from 4 microarrays with the batch effect
Supplementary Table 4C - PDF file - 111K, Significance of methylation in prostate cancer tissue samples from 4 microarrays with the batch effect
Supplementary Table 4D - PDF file - 267K, Significance of methylation in prostate cancer tissue samples from 4 microarrays with the batch effect
Supplementary Table 5 - PDF file - 172K, Prostate cancer tissues used for methylation analysis by pyrosequencing
Supplementary Table 3 - XLS file - 43K, Most significantly methylated genes associated with prostate cancer recurrence from the Cox multivariate model
Supplementary Table 2 - XLS file - 34K, Most significantly methylated genes associated with prostate cancer recurrence from the Cox univariate analysis