Usage

Localize and undrift

General description

To localize and undrift the raw ome.tif files the following steps are performed:

  1. An automated minimal net-gradient detection is performed to define all valid spots
  2. A 2D Gaussian least square fit is performed for every spot using picasso.localize
  3. The resulting localizations are undrifted in segments by RCC as described in picasso.render

How to use

The localize_undrift notebook guides through the usage of the picasso_addon.localize.main() function to obtain localized and undrifted localizations lists from raw ome.tif files. If you prefer normal pyhton scripts (e.g. for use in spyder) you can find it here.

Autopick

General description

To automatically define localization clusters in the rendered localizations (_render.hdf5) we go from a pointillistic to a pixel based presentation. We can then use a modified version of the spot finding algorithm of picasso.localize. The following steps are performed:

  1. The localization list is rendered to a subpixel image given by oversampling (see picasso.render).
  2. Every subpixel value now stands for the number of localizations within each subpixel area.
  3. The spot detection function of picasso.localize is employed to define spots (boxes) in the rendered localization image.
  4. The number of all localizations within these boxes is calculated and we only consider boxes above a certain threshold min_n_locs.
  5. We calculate the center of mass (i.e. localizations) within the remaining boxes. This gives us the pick center coordinates.
  6. The pick center coordinates are saved as _autopick.yaml.
  7. We employ a KDtree to get all localizations with a distance less than pick_diameter/2 to the pick center coordinates to obtain the picks.
  8. Last, we give every localization an ID group corresponding to its pick identity. Localizations not corresponding to a pick are disregarded.
  9. The result is saved as _picked.hdf5.

How to use

The autopick notebook guides through the usage of the picasso_addon.autopick.main() function to obtain localization clusters above a certain localization threshold from *_render.hdf5 localization lists. If you prefer normal python scripts (e.g. for use in spyder) you can find it here.

Detailed info

Data structure of localization lists

Please have a look at this notebook to learn how to load picassos’ localization lists and which observables it contains.

Automated minimal net-gradient detection

We will insert a notebook here to explain what is happening in detail, please be patient.