Good Afternoon Joseph: I believe what you are seeing is the effect of the conversion procedure from your original bedGraph data into the genome graph data points. The conversion is a weighted average into 10,000 base windows. The Genome graphs are line graphs through data points at every 10,000 bases. The original bedGraph data on the genome browser are bar graphs over intervals of bases. The graphing paradigms are fundamentally different.
Looking at the first 7 data points in one of your examples. For example, the following four points in the original bedGraph data: chr1 44006 44066 -0.12926 chr1 46023 46083 -0.0995345 chr1 47390 47601 -0.0876306 chr1 48251 48311 -0.0034941 Becomes a weighted average over 10,000 bases: -0.003243 = ((60 * -0.12926) + (60 * -0.0995345) + (211 * -0.0876306) + (60 * -0.0034941)) / 10000 And the next three points: chr1 80315 80375 -0.0104523 chr1 83080 83140 -0.0377893 chr1 84485 84545 -0.0539718 Weighted average: -0.000613 = ((60 * -0.0104523) + (60 * -0.0377893) + (60 * -0.0539718)) / 10000 Produces Genome Graphs data points: chr1 5000 0 chr1 15000 0 chr1 25000 0 chr1 35000 0 chr1 45000 -0.003243 chr1 55000 0 chr1 65000 0 chr1 75000 0 chr1 85000 -0.000613 chr1 95000 0 ... etc ... Given such a conversion, the resulting data can appear much different than the original input data. If you have a better windowing conversion method, you can construct directly the genome graphs data view and upload it to genome graphs. --Hiram > Genome graph doesn't seem to be displaying my data correctly. I see mostly > positive lines on the graph but when I look at my data there are many > negative > values. Does Genome Graphs do something to the negative values? Two datasets > that should display as exact opposites are far from what I expected. > To get this problem I click on "add custom tracks" uploading a custom track > type=bedGraph. Then in Genome Graphs I import the custom track using > zebrafish > zv7 settings. I can give you the file I am uploading if that helps. > Thank you for your help so far. I hope I can get this working because Genome > Browser has been a very useful tool for me so far. > Joseph _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
