Merge remote-tracking branch 'origin/master' into boli

docs/conf.py

@@ -23,15 +23,15 @@
 # -- Project information -----------------------------------------------------
 
 project = "Pegasus"
-copyright = "2023 Genentech, Inc. All rights reserved."
+copyright = "2024 Genentech, Inc. All rights reserved."
 author = (
     "Yiming Yang, Joshua Gould and Bo Li"
 )
 
 # The short X.Y version
-version = "1.8"
+version = "1.9"
 # The full version, including alpha/beta/rc tags
-release = "1.8.1"
+release = "1.9.0"
 
 
 # -- General configuration ---------------------------------------------------

docs/index.rst

@@ -10,7 +10,7 @@
 Release Highlights in Current Stable
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. include:: release_notes/version_1_8.rst
+.. include:: release_notes/version_1_9.rst
 
 .. toctree::
    :maxdepth: 1

docs/release_notes.rst

@@ -6,6 +6,11 @@ Release Notes
 .. note::
     Also see the release notes of `PegasusIO <https://pegasusio.readthedocs.io/en/stable/release_notes.html>`__.
 
+Version 1.9
+~~~~~~~~~~~~~
+
+.. include:: release_notes/version_1_9.rst
+
 Version 1.8
 ~~~~~~~~~~~~~
 

docs/release_notes/version_1_9.rst

@@ -0,0 +1,15 @@
+1.9.0 :small:`January 19, 2024`
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+**New Feature and Improvement**
+
+* ``calculate_z_score`` works with sparse count matrix. [PR `276 <https://github.com/lilab-bcb/pegasus/pull/276>`_ Thanks to `Jayaram Kancherla <https://github.com/jkanche>`_]
+* Plotting functions (``scatter``, ``dotplot``, ``violin``, ``heatmap``) now give warnings on genes/attributes not existing in the data, and skip them in the plots.
+* Improve ``heatmap``:
+
+  * Add ``show_sample_name`` parameter for cases of pseudo-bulk data, nanoString DSP data, etc.
+  * Use Scipy's linkage (``scipy.cluster.hierarchy.linkage``) for dendrograms to use its optimal ordering feature for better results (see ``groupby_optimal_ordering`` parameter).
+
+* Update human lung and mouse immune markers used by ``infer_cell_types`` function.
+* ``run_harmony`` can accept multiple attributes to be the batch key, by providing a list of attribute names to its ``batch`` parameter.
+* Expose ``online_batch_size`` parameter in ``nmf`` and ``integrative_nmf`` functions.

docs/requirements.txt

@@ -14,7 +14,6 @@ natsort
 joblib
 psutil
 numba
-importlib_metadata; python_version < '3.8'
 umap-learn
 forceatlas2-python
 pyarrow

pegasus/tools/batch_correction.py

@@ -1,7 +1,7 @@
 import time
 import numpy as np
 import pandas as pd
-from typing import Union
+from typing import Union, List
 from pegasusio import UnimodalData, MultimodalData
 
 from pegasus.tools import select_features, X_from_rep, check_batch_key
@@ -16,7 +16,7 @@
 @timer(logger=logger)
 def run_harmony(
     data: Union[MultimodalData, UnimodalData],
-    batch: str = "Channel",
+    batch: Union[str, List[str]] = "Channel",
     rep: str = "pca",
     n_comps: int = None,
     n_jobs: int = -1,
@@ -34,8 +34,8 @@ def run_harmony(
     data: ``MultimodalData``.
         Annotated data matrix with rows for cells and columns for genes.
 
-    batch: ``str``, optional, default: ``"Channel"``.
-        Which attribute in data.obs field represents batches, default is "Channel".
+    batch: ``str`` or ``List[str]``, optional, default: ``"Channel"``.
+        Which attribute in data.obs field represents batches, default is "Channel". If using multiple attributes, specify their names in a list.
 
     rep: ``str``, optional, default: ``"pca"``.
         Which representation to use as input of Harmony, default is PCA.
@@ -54,7 +54,7 @@ def run_harmony(
 
     use_gpu: ``bool``, optional, default: ``False``.
         If ``True``, use GPU if available. Otherwise, use CPU only.
-        
+
     max_iter_harmony: ``int``, optional, default: ``10``.
         Maximum iterations on running Harmony if not converged.
 

pegasus/tools/utils.py

@@ -157,19 +157,24 @@ def simulate_doublets(X: Union[csr_matrix, np.ndarray], sim_doublet_ratio: float
     return results, doublet_indices
 
 
-def check_batch_key(data: Union[MultimodalData, UnimodalData], batch: str, warning_msg: str) -> bool:
+def check_batch_key(data: Union[MultimodalData, UnimodalData], batch: Union[str, List[str]], warning_msg: str) -> bool:
     if batch is None:
         return False
 
-    if batch not in data.obs:
-        logger.warning(f"Batch key {batch} does not exist. {warning_msg}")
-        return False
-    else:
-        if not is_categorical_dtype(data.obs[batch]):
-            data.obs[batch] = pd.Categorical(data.obs[batch].values)
-        if data.obs[batch].cat.categories.size == 1:
-            logger.warning(f"Batch key {batch} only contains one batch. {warning_msg}")
+    if isinstance(batch, str):
+        batch = [batch]
+
+    for bat in batch:
+        if bat not in data.obs:
+            logger.warning(f"Batch key {bat} does not exist. {warning_msg}")
             return False
+        else:
+            if not is_categorical_dtype(data.obs[bat]):
+                data.obs[bat] = pd.Categorical(data.obs[bat].values)
+            if data.obs[bat].cat.categories.size == 1:
+                logger.warning(f"Batch key {bat} only contains one batch. {warning_msg}")
+                return False
+
     return True
 
 

requirements.txt

@@ -15,7 +15,7 @@ natsort
 numba
 numpy
 pandas>=1.2.0
-pegasusio>=0.5.1
+pegasusio>=0.9.0
 pybind11
 scikit-learn>=0.23.2
 scikit-misc

setup.py

@@ -59,7 +59,7 @@
         scvi=["scvi-tools"],
         all=["fitsne", "louvain", "scanorama", "torch", "harmony-pytorch", "nmf-torch", "rpy2", "forceatlas2-python", "scvi-tools"]
     ),
-    python_requires="~=3.7",
+    python_requires="~=3.8",
     package_data={
         "pegasus.annotate_cluster": [
             "human_immune_cell_markers.json",