comparison hicCorrectMatrix.xml @ 10:bfa1c014f64a draft

planemo upload for repository https://github.com/maxplanck-ie/HiCExplorer/tree/master/galaxy/wrapper/ commit dddc0b9035b8edadfd45d74b01aeca245c2725d7

9:ac80bd0a96ca

------

Diagnostic plot
_______________

The diagnostic plot consists of a bar plot of the contacts coverage per bins size together with the
modified z-score based on the Median Absolute Deviation (MAD) method.

See Boris Iglewicz and David Hoaglin 1993, Volume 16:
How to Detect and Handle Outliers The ASQC Basic References in Quality Control: Statistical Techniques,
Edward F. Mykytka, Ph.D., Editor.

Using this diagnostic plot, a user can decide if values
with a too low (and/or too high) number of contacts in respect to their genomic distance should
be removed from the data before the correction applies.

Moreover, the shown distribution should be a Gaussian bell. If it doesn’t follow a Gaussian distribution
this is an indicator that the used data is of bad quality or that the used contact matrix
is maybe not the one that should be used. It can happen that users select for example a merge
matrix with a lower resolution that was previously needed for plotting. In such cases the
diagnostic plot helps to detect this and prevent the user from running the analysis on a wrong dataset.


.. image:: $PATH_TO_IMAGES/diagnostic_plot.png
    :width: 50%

On the example plot above, a user can then use the lower threshold defined by the MAD method (black bold bar), or define its own threshold based on the contacts distribution.

Correct
_______

Run the iterative correction and outputs the corrected matrix. This matrix can then be used with all downstream analysis tools such as ``hicPlotMatrix``, ``hicPlotTADs``, ``hicPlotViewpoint``, ``hicAggregateContacts`` for **visualization of Hi-C data**, ``hicCorrelate``, ``hicPlotDistVsCounts``, ``hicTransform``, ``hicFindTADs``, ``hicPCA`` **for data and scores computation on Hi-C data**.

10:bfa1c014f64a

------

Diagnostic plot
_______________


The goal of the diagnostic plot is to help the user decide on a cutoff threshold that will ignore Hi-C matrix
bins with few reads assigned to them. The plot is a histogram of the total number of Hi-C reads per matrix bin.
A secondary scale based on the mean absolute deviation score, is shown on top of the figure.
This secondary scale aims to offer 'normalized' values that are comparable across samples
independently of the sequencing depth and the fraction of usable Hi-C reads. In all samples that we have studied,
the histogram follows a bimodal distribution where the first peak is for bins with zero reads which usually occur
at repetitive regions. Other low scoring bins tend to be close to repetitive regions.
Also, low scoring bins can be caused by absence of a restriction site in the bin or because the restriction
site is present but the restriction enzyme did not cut. The valley between the two peaks in the
histogram is set by default as cutoff threshold.
However, it is important to revise this as in some cases the selected value could not be correct. 


.. image:: $PATH_TO_IMAGES/diagnostic_plot.png
    :width: 50%

On the example plot above, a user can then use the lower threshold defined by the Median Absolute Deviation (MAD) method (black bold bar), or define its own threshold based on the contacts distribution.

Correct
_______

Run the iterative correction and outputs the corrected matrix. This matrix can then be used with all downstream analysis tools such as ``hicPlotMatrix``, ``hicPlotTADs``, ``hicPlotViewpoint``, ``hicAggregateContacts`` for **visualization of Hi-C data**, ``hicCorrelate``, ``hicPlotDistVsCounts``, ``hicTransform``, ``hicFindTADs``, ``hicPCA`` **for data and scores computation on Hi-C data**.