What's new with the ChIPper?

This document lists changes made to the ChIPper website over time. If it's been a while since the last time you visited, you might want to scan the list to see if there are any changes relevant to you.

November 2005
- Published at Genome Biology:
  Chipper: discovering transcription-factor targets from chromatin immunoprecipitation microarrays using variance stabilization Francis D Gibbons, Markus Proft, Kevin Struhl, Frederick P Roth Genome Biology 2005, 6:R96 (1 November 2005) PDF
August 2005
- Submitted for publication at Genome Biology in March, the (extensive) revisions are under consideration there now.
- This version has been brought up to date with the current version of the R code, which has changed significantly since this website was re-built. Look and feel remain the same.
August 2004
- Now uses likelihood maximization to estimate parameters for normal fitted to 'lefthand' edge of unbound population (Δh distribution).
July 2004
- Internet Explorer 5.0 on Mac OS 9.x does not understand HTML <button> tag. Reworked code to use older <input> tag to same effect.
June 2004

Complete rewrite from scratch. If you used it before, you won't recognize it now. Among the changes you'll see are the following:

You must upload files in a very simple format: three columns, containing some kind of unique identifier, the median foreground IP data, and the median foreground control ('mock') data. Do not perform any background subtraction or normalization on the data. That's all built into the algorithm used here (specifically by use of the vsn package). The columns must be labelled 'Promoter', 'ip', and 'control. Here's an example:

Promoter	ip	control
iYALCdelta2	11393.0	10345.0
iYBRCdelta11	118659.0	117221.0
iYDRWdelta23	131070.0	131070.0
iYELCdelta4	125494.0	90662.0
iYFLWdelta1	29450.0	23945.0
iCEN5	45021.0	38191.0
iSNR13	6342.0	7645.0
iSNR189	24733.0	26942.0
iSNR190	17839.0	16047.0
iSNR44	17286.0	14906.0
iSNR45	22947.0	23018.0

If you're wondering about the form of the identifiers, they're composed of three parts, which commonly appear in the raw data files produced by a microarray experiment: intergenic region, ORF, gene name. Neither is sufficient on its own. A given intergenic region may be a potential promoter of two divergently transcribed genes. Gene names and ORFs are inherently changeable, as annotation changes.

You may upload several files, containing 'replicates' (the same experiment done on different chips) and/or 'duplicates' (the same experiment with each intergenic region duplicated on the chip).
The method for computing P-values for each experiment has changed completely. We believe that this is better than anything we have tried before.
Data are returned ranked according to increasing P-value. For each intergenic region, the P-value combines the results from all uploaded datasets.
Formatting a table with thousands of lines is hard work for most computers. In order to get your results more quickly, you may choose to limit the size of the ranked list that's returned to you when the computation finishes. You can choose to see a certain number of the top-ranked hits, or only those that satisfy a certain level of significance. Of course, if you really want it, you can have the entire list.