This track displays activity levels of 643 putative promoter fragments in the ENCODE regions, based on high-throughput transient transfection luciferase reporter assays. The activity of each putative promoter is indicated by color, ranging from black (no activity) to red (strong activity). Each of the fragments was tested in a panel of 16 cell lines:
Cell Line Classification Isolated From AGS gastric adenocarcinoma stomach BE(2)-C neuroblastoma brain (metastatic, from bone marrow) T98G (CRL-1690) glioblastoma brain G-402 renal leiomyoblastoma kidney HCT 116 colorectal carcinoma colon HMCB melanoma skin HT-1080 fibrosarcoma connective tissue SK-N-SH (HTB-11) neuroblastoma brain (metastatic, from bone marrow) HeLa adenocarcinoma cervix HepG2 hepatocellular carcinoma liver JEG-3 choriocarcinoma placenta MG-63 osteosarcoma bone MRC-5 fibroblast lung PANC-1 epithelioid carcinoma pancreas (duct) SNU-182 hepatocellular carcinoma liver U-87 MG glioblastoma-astrocytoma brain
Promoters in the ENCODE region were predicted using a variation on methods previously described (Trinklein et al., 2003, Trinklein et al., 2004). Using BLAT alignments of human cDNAs in Genbank to the genome, those with at least one bp of exon overlap were merged, generating gene models. The transcription start sites were predicted by assigning the 5' end of each gene model as one transcription start site and alternative 5' ends that were at least 500 bp downstream and supported by full-length cDNAs as other start sites. Promoters were defined as the regions approximately 600 bp upstream and 100 bp downstream of each transcription start site.
Primer3 was used to pick primers yielding approximately 500 bp amplicons containing the predicted transcription start site. Each fragment of DNA represented in this track was cloned into a luciferase reporter vector (pGL3-Basic, Promega) using the BD Clontech Infusion Cloning System. The Dual Luciferase system (Promega) was used to co-transfect the experimental DNA along with a control plasmid expressing Renilla - to control for variation in transcription efficiency - in 96-well format into one of the sixteen cell types using FuGENE Transfection Reagent (Roche). Each transfection was done in duplicate.
Data are reported as normalized and log2 transformed averages of the Luciferase/Renilla ratio. This normalization was based on the activity of 102 random genomic fragments (negative controls) derived from exons and intergenic regions. Such a normalization allows for a meaningful comparison between cell types. The average log transformed Luciferase/Renilla ratio was scaled linearly to create a score where the maximum value is 1000 and the minimum value is 0. This score is arbitrary and for visualization purposes only; the raw ratio values should be used for all analyses.
Data were verified by repeating the preparation and measurement of 48 random fragments. No significant variation between the two preparations was detected.
A spreadsheet containing the negative control data can be downloaded here.
This work was done in collaboration at the Myers Lab at Stanford University (now at HudsonAlpha Institute for Biotechnology). The following people contributed: Sara J. Cooper, Nathan D. Trinklein, Elizabeth D. Anton, Loan Nguyen, and Richard M. Myers.
Trinklein ND, Aldred SJ, Saldanha AJ, Myers RM. Identification and functional analysis of human transcriptional promoters. Genome Res. 2003 Feb;13(2):308-12.
Trinklein ND, Aldred SF, Hartman SJ, Schroeder DI, Otillar RP, Myers RM. An abundance of bidirectional promoters in the human genome. Genome Res. 2004 Jan;14(1):62-6.