Tutorial should be a first pass now

docs/tutorial.rst

@@ -5,18 +5,108 @@ the package to annotate your single cell dataset based on one or more cell atlas
 the end of the tutorial, you should be able to navigate the other docs yourself (see
 :doc:`/api`).
 
-Starting point: atlas landmarks
+Start off: atlas landmarks
 --------------------------------
 To transfer cell types from an atlas, you need a few cells or averages for each cell type
 within the atlas. We call these **atlas landmarks**. To keep things simple, in this tutorial
 we use precomputed landmarks from our `sister project <https://iosonofabio.github.io/atlas_landmarks/>`_.
-Alternatively, you can use a custom atlas with its landmarks, see the api_ documentation of
+Alternatively, you can use a custom atlas with its landmarks, see the :doc:`api` documentation of
 classes :class:`.Averages` and :class:`.Subsample` for that advanced usage.
 
+For this tutorial, we will use the atlas `Darmanis_2015 <https://www.pnas.org/content/112/23/7285>`_.
 
-Preparing your single cell dataset
+
+Prepare your single cell dataset
 --------------------------------------
-TODO
+Then we need to prepare the new single cell dataset to annotate. `northstar` accepts two formats:
+
+1. a `pandas.DataFrame` with features/genes as rows and cells as columns.
+
+2. an `anndata.Anndata` object. Turns out `Anndata <https://anndata.readthedocs.io/en/stable/>`_
+   enforces the opposite convention, so the rows/observation_names must be the cells and the
+   columns/variable_names must be the features/genes.
+
+.. note::
+   `northstar` will take the intersection of your features names and the atlas features to
+   assign cell types. Most atlases use gene names instead of EnsemblIDs or other names, so
+   make sure you do the same.
+
+For this tutorial, we will use pandas, so let's load the data e.g. from a CSV file:
+
+.. code-block:: python
+
+  # Assume the CSV already has genes as rows
+  dataset = pd.read_csv('mydataset.csv', sep=',', index_col=0)
+
+If the dataset is in a `loom <http://loompy.org/>`_ file, we have to set the index and columns:
+
+.. code-block:: python
+
+  import loompy
+
+  # Assume the gene name is in an attribute called GeneName and the cell name in CellID
+  with loompy.connect('mydataset.loom') as ds:
+      dataset = pd.DataFrame(
+          data=ds[:, :],
+          index=ds.ra['GeneName'],
+          columns=ds.ca['CellID'],
+          )
+
+Create an instance of northstar
+-------------------------------
+Let's assume you want to use the `.Subsample` class. You can create an instance of the class
+as follows:
+
+.. code-block:: python
+
+  import northstar
+
+  model = northstar.Subsample(
+      atlas='Darmanis_2015',
+      new_data=dataset,
+      )
+
+Call the cell classifier
+------------------------
+This is where the actual computations happen. A `northstar` instance can be called like a
+normal function:
+
+.. code-block:: python
+
+  model()
+
+Extract the result
+------------------
+The result of the cell type assignment can be extracted by the following command:
+
+.. code-block:: python
+
+  cell_types = model.membership
+
+This is a numpy array with the same length and order as your cells.
+
+Optional: closest atlas cell type
+---------------------------------
+Sometimes you get novel clusters that do not match any atlas cell type. To start identifying
+those clusters, you can ask `northstar` what known atlas cell type they are most similar to.
+Here's the code to do that:
+
+.. code-block:: python
+
+  closest_cell_types = model.estimate_closest_atlas_cell_type()
+
+The output is a `pandas.Series` with the novel clusters as index and the closest atlas cell
+types as values.
+
+Where to go from here
+---------------------
+This concludes this short tutorial that showcases the main usage of `northstar`. You can
+browse the :doc:`/api` page for more detailed information such as how to combine atlases,
+specify only certain cell types within one or multiple atlases, and use custom atlases.
+
+We hope `northstar` helps you understand your tissue sample and do not hesitate to open an
+`issue on github <https://github.com/iosonofabio/northstar/issues>`_ if you have trouble.
+If `northstar` was useful for a publication, please consider citing us at `PAPER MISSING`.
 
 .. toctree::
    :hidden: