Hi Carly, While we don't offer one particular tool to retrieve the information you're requesting, one suggestion is to create a custom track containing transcription start positions + and - 500bp, and intersect it with the following tables in the Table Browser: wgEncodeUwHistoneHepg2H3k27me3StdPkRep1 wgEncodeUwHistoneHepg2H3k27me3StdPkRep2 wgEncodeUwHistoneBjH3k27me3StdPkRep1 wgEncodeUwHistoneBjH3k27me3StdPkRep2
To find the actual transcription start positions, you will need to query the knownCanonical table in the Table Browser (to get only one isoform). You will also need to filter by strand on the knownGene table by clicking on the 'create' button next to 'filter:', and setting 'strand does match +' (for the positive strand). You can retrieve the chrom and chromStart by choosing 'selected fields from primary and related tables' as the output format, and selecting those fields on the following page. Then you will need to add 1 to chromStart to get the 3rd field of the BED file. You can do this using excel or if you are familiar with bash (or tc shell) you could use that too. You should then have a BED file looking like this example (for the + strand): track name="+ strand canonical starts" chr1 11873 11874 chr1 69090 69091 chr1 321083 321084 chr1 321145 321146 chr1 322036 322037 chr1 367658 367659 chr1 420205 420206 chr1 566461 566462 chr1 568148 568149 Then, you could use the Table Browser (or Excel or a script) to retrieve a BED file with +/- 500bp. For the negative strand genes, you would need to query the Table Browser for chromEnd. For more information on the Table Browser, Custom Tracks or BED file format, please see the following help pages: http://genome.ucsc.edu/goldenPath/help/hgTablesHelp.html http://genome.ucsc.edu/goldenPath/help/customTrack.html http://genome.ucsc.edu/FAQ/FAQformat.html#format1 Please contact us again at [email protected] if you have any further questions. --- Luvina Guruvadoo UCSC Genome Bioinformatics Group On 1/19/2012 2:36 PM, Carly Hom wrote: > Hi my name is Carly Hom and I am an undergraduate student researcher at > Arizona State University working with Dr. Karmella Haynes. In my current > lab I am using Synthetic Biology and Bioinformatics to investigate reliable > and predictable reactivation of dormant genes that can help treat cancer > and enable tissue re-growth. By determining which silenced genes will > switch to an active state in osteosarcoma cells, with the presence of the > synthetic transcription factor PC-TF, my work will establish a > comprehensive method for predicting the effect of rationally designed > protein-based drugs. Pc-TF, a synthetic transcription factor developed by > Dr. Haynes, regulates cell states by binding the repressive > trimethyl-histone H3 lysine 27 signal (H3K27me3) and switching silenced > genes to an active state in osteosarcoma cells. Since a comprehensive ChIP > map is not available for osteosarcoma, I will be identifying genes > associated with H3K27me3 in liver (HepG2) and fibroblast (BJ) cell lines. > Overall, I will need to collect about 1000 genes from the ENCODE database > that show a significant enough H3K27me3 signal at the promoter of the gene. > I have already figured out how to project only information from the HepG2 > and BJ cell lines in relation to H3K27me3, but by just clicking to move > through the cell line to find genes will take entirely too long and can > cause me to miss important genes. At the request of Dr. Haynes I am asking > if ENCODE has some sort of filter program that will provide a list of genes > where the promoter site shows a high level of the H3K27me3 histone > methylation. I will need it to be able to find the beginning of the gene's > promoter on the UCSC Genome Browser and then show about 500bps to the left > and 500bps to the right of the promoter . Ultimately, I want to be able to > navigate through the genes in this cell line that show a significant enough > H3K27me3 signal at the promoter (everything else with a low H3K27me3 signal > I do not care about). If you could get back to me on whether this is even > possible to do within the Genome Browser, and if yes, how I would be able > to do this that would be great. Thank you! > > - Carly Hom > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
