Good Afternoon Basel: I can not tell exactly what is happening to your data display without understanding the original data. Do you have the wigToBigWig kent command to turn your original wiggle ascii data into bigWig file formats ? You may be able to see your data better with bigWig file formats. The older wiggle encoding is a lossy data encoding format and can distort the display of data depending upon the characteristics and density of the data.
You should also keep in mind, even with an accurate data encoding scheme as in bigWig, when the graph is constructed on a limited pixel sized display, the multiple data points in some number of underlying bases of the genome will combine together to form the data value at a single pixel of the graph. Depending upon the size of the viewing window and the density of the data, this often has the effect of smoothing out the data to a mean value of the many underlying data values. You can also control how the data is displayed by changing the parameters of the track behavior in the track controls. You can select to view the minimum, maximum, or mean of the data values that combine into a single pixel of the graph. I would be curious to take a look at one of your original wiggle ascii data files before it was submitted as a custom track. If you can privately send me a URL to one of your data files, I'll take a look. --Hiram Baghal, Basel (NIH/NIAAA) [F] wrote: > Hello, > > I have encountered a problem when visualizing my custom Wig tracks. I first > noticed that after loading the files, some of the tracks were missing peaks > for very long stretched (sometimes an entire arm of a chromosome) while the > files themselves do contain the data. After further investigation, I realized > that the problematic tracks were those containing very large peaks at near > the centromeric boundaries. It appears that when a large peak (>10,000) > occurs all the bins to the left and some to the right disappear or are > distorted. I have search the site (mainly the WIG Format page) for any > guidelines on a maximum number of counts per bin but was unable to find any > such information. > > Is there a limit to the peak size beyond which the track can no longer be > properly visualized? > > I have prepared an example of the problem I have been experiencing. I have > loaded 5 Wig tracks showing my original data and 4 tests in which I gradually > decrease the limit on the maximum peak permitted (the data is all the same > except that peaks above a certain value are decreased to a set threshold). > Test1 does not allow a value above 100,000 , test2 - 50,000, test3 - 25,000, > and test4 - 10,000. What I observe is that with each decrease the "real" data > is better resolved and less distorted. > > My Sessions: Example1 shows Chromosome 10 in its entirety (the p-arm is > distorted) and Example2 zooms in on a region in the p-arm that shows the > problem more clearly. > > Here is a UCSC browser session I'd like to share with you: > http://genome.ucsc.edu/cgi-bin/hgTracks?hgS_doOtherUser=submit&hgS_otherUserName=Baghalb&hgS_otherUserSessionName=WigTrack_problem_example1 > > Here is a UCSC browser session I'd like to share with you: > http://genome.ucsc.edu/cgi-bin/hgTracks?hgS_doOtherUser=submit&hgS_otherUserName=Baghalb&hgS_otherUserSessionName=WigTrack_problem_example2 > Sincere thanks, > > Basel Baghal _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
