Pylake 1.4.0
============

.. only:: html

Pylake `v1.4.0` has been released with new features and improvements to existing analyses. Here’s some of the highlights:

Hidden Markov Models
--------------------

Hidden Markov Models (HMMs) are often used for analyzing data that shows transitions among discrete
states. Now with just a few lines of code you can fit any channel data and view the results.
If you're interested in kinetics, the model also generates data that can be used with
:class:`~lumicks.pylake.DwelltimeModel` to extract state lifetimes.

Check out the :ref:`hmm-section` tutorial and the
:class:`~lumicks.pylake.HiddenMarkovModel` API page for more information.

.. figure:: hmm_hairpin.png

   HMM analysis of a tethered DNA hairpin held at three different bead separations.

Automatic bead cropping
-----------------------

Added :meth:`~lumicks.pylake.kymo.Kymo.estimate_bead_edges()` and :meth:`~lumicks.pylake.kymo.Kymo.crop_beads()` to quickly crop the beads out of a kymograph using an estimate of the bead edges. This can help when batch processing kymographs.

.. figure:: bead_edges.png

Filter customization kymotracking
---------------------------------

We added the option to customize the filters applied prior to peak detection to :func:`~lumicks.pylake.track_greedy`.
To do this, we added two additional parameters:

- `filter_width` allows customizing the filter applied prior to detection.
- `adjacency_filter` applies a filter on the detected peaks, removing any fluorescent peaks that do not have a detected peak in an adjacent frame.

This allows using lower thresholds, while keeping false detections in check.

.. figure:: tracking_comparison.png

.. figure:: tracking_comparison_threshold.png

Other changes
-------------

In addition, this release contains several other bug-fixes and improvements.
For a full list of all the changes, please refer to the full :doc:`changelog</changelog>`.

Happy Pylake-ing!