metaseq.integration.chipseq.Chipseq

class metaseq.integration.chipseq.Chipseq(ip_bam, control_bam, dbfn=None)

Bases: object

Class for visualizing and interactively exploring ChIP-seq data.

Needs two BAM files (for IP and control) and a gffutils database filename in order to display gene models.

Typical usage is to create a normalized array of signal over each feature with the diff_array method, and then plot with the plot method.

The resulting figure has the matrix as a heatmap, the average signal over features, and a panel with points that can be zoomed and clicked, spawning a mini-browser window for the corresponding feature.

Configuration can be done by adjusting the following attributes after creating a Chipseq instance:

• _strip_kwargs (the style for the dots in the left panel)
• browser_plotting_kwargs (style of signal lines in the mini-browser)

Methods

callback(event)	Callback function to spawn a mini-browser when a feature is clicked.
diff_array(features[, force, func, ...])	Scales the control and IP data to million mapped reads, then subtracts scaled control from scaled IP, applies func(diffed) to the diffed array, and finally sets self.diffed_array to be the result.
plot(x[, row_order, imshow_kwargs, strip])	Plot the scaled ChIP-seq data.

Set up a Chipseq object.

Parameters:

• ip_bam – filename of BAM file for ChIP data
• control_bam – filename of BAM file for control data
• dbfn – filename of gffutils database

Methods

callback(event)	Callback function to spawn a mini-browser when a feature is clicked.
diff_array(features[, force, func, ...])	Scales the control and IP data to million mapped reads, then subtracts scaled control from scaled IP, applies func(diffed) to the diffed array, and finally sets self.diffed_array to be the result.
plot(x[, row_order, imshow_kwargs, strip])	Plot the scaled ChIP-seq data.

Methods

__init__(ip_bam, control_bam[, dbfn])	Set up a Chipseq object.
callback(event)	Callback function to spawn a mini-browser when a feature is clicked.
diff_array(features[, force, func, ...])	Scales the control and IP data to million mapped reads, then subtracts scaled control from scaled IP, applies func(diffed) to the diffed array, and finally sets self.diffed_array to be the result.
plot(x[, row_order, imshow_kwargs, strip])	Plot the scaled ChIP-seq data.

__init__(ip_bam, control_bam, dbfn=None)

Set up a Chipseq object.

Parameters:

• ip_bam – filename of BAM file for ChIP data
• control_bam – filename of BAM file for control data
• dbfn – filename of gffutils database

callback(event)

Callback function to spawn a mini-browser when a feature is clicked.

diff_array(features, force=True, func=None, array_kwargs={}, cache=None)

Scales the control and IP data to million mapped reads, then subtracts scaled control from scaled IP, applies func(diffed) to the diffed array, and finally sets self.diffed_array to be the result.

Arrays self.ip and self.control are set as well, and if force=False, then previously-created arrays will be used instead of re-calculating new ones. This is useful if you want to easily try multiple func functions without having to re-calculate the data.

Another side-effect is that self.features is set so that it can be accesed by other methods.

Parameters:

• features – a list of pybedtools.Interval objects
• array_kwargs – extra keyword args passed to genomic_signal.array; typically this will include bins, processes, and chunksize arguments.
• func – a function to apply to the diffed arrays. By default this is metaseq.plotutils.nice_log(); another option might be lambda x: x, or lambda x: 1e6*x
• force – Force a re-calculation of the arrays; otherwise uses cached values

plot(x, row_order=None, imshow_kwargs=None, strip=True)

Plot the scaled ChIP-seq data.

Parameters:

• x – X-axis to use (e.g, for TSS +/- 1kb with 100 bins, this would be np.linspace(-1000, 1000, 100))
• row_order – Array-like object containing row order – typically the result of an np.argsort call.
• strip – Include axes along the left side with points that can be clicked to spawn a minibrowser for that feature.