User:Karmella Haynes/Notebook/PcTF Genomics/2013/01/07

01/07/13

• Array Star analysis

Array Star analysis

1. Open Array Star. Windows only. This can be run on Parallels from a Mac.
2. Click "Start Chip-Seq project..."
3. Add Experiments to Import: Click [Add File..]
4. Select a ###.bam file from the ChiP seq data folder (shared desktop). Click [Next >].
5. Create binding proteins: Click [Create Binding Protein]. Binding protein label = the name of the protein that was pulled down by the antibody via ChIP. Binding site type = Unknown. Click [OK].
6. Assign Binding Proteins to Experiments: Protein label = the appropriate binding protein (created in previous step). Click [Next >].
7. Setup Preprocessing: Desired preprocessing = QSeq; Normalization method = none; Sequence files against which to quantify = browse to Desktop > Human Genome Ref > select all NC...gbk files.
   1. Select "Use features of type(s)" and set to "gene".
   2. Genome filtering = Discover peaks in the entire genome.
   3. Configure Advanced Options: under General, Graphs > Create Wiggle (.wig). Set Location to ChIPseq Wig Files on desktop.
   4. Configure Advanced Options: under General, Alignment Files > Create BED Alignment (.bed). Set Location to ChIPseq BED Files on desktop.
   5. Configure Advanced Options: under Peak Detection, is set to Qseq Peak Finder by default. Leave this as-is for now.
   6. Click [Next >]. Wait a while
8. Setup Attributes and Replicates: Show only grouping attributes (default). Click [Finish].
9. Window with Peak Table and other tabs should appear. Finished! Explore the data.

PcTF vs. H3K27me3

• Add Experiments to Import: added 1_aln_sorted.bam (PcTF), 2_aln_sorted.bam (PcTF mock), 6_aln_sorted.bam (H3K27me3), and 7_aln_sorted.bam (H3K27me3 mock)
• Create binding proteins: created "PcTF" for 1_aln_sorted, and "H3K27me3" for 6_aln_sorted.
• Set control 2_ and 7_aln_sorted as "yes" for "Is control?"
• Set 2_aln_sorted as control for 1_aln_sorted. Set 7_aln_sorted as control for 6_aln_sorted.
  • This analysis required too much disk space
• Started over, but only used chromosomes 1-3 as reference templates (trying a few at a time).
  • This worked

Results/ Notes:

• Peak Table: shows very few hits for PcTF, many more for H3K27me3, no overlap!
• Display wig file on UCSC Genome browser online - add file from "ChIPseq Wig Files" folder as a custom track; wig file shows raw data histogram
  • 1_aln_sorted.wig, 6_aln_sorted.wig
• Display BED file on UCSC Genome browser online - add file from "ChIPseq BED Files" folder as a custom track; BED file shows
  • 1_aln_sorted.bed, 6_aln_sorted.bed