Release 0.1.2. Minor updates / bug fixes + Adding containers.

Author: shz9

Diff for: CHANGELOG.md

@@ -5,6 +5,20 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.1.2] - 2024-04-24
+
+### Changed
+
+- Fixed `manhattan` plot implementation to support various new features.
+- Added a warning when accessing `csr_matrix` property of `LDMatrix` when it hasn't been loaded
+previously.
+
+### Added
+
+- `reset_mask` method for magenpy `LDMatrix`.
+- `Dockerfile`s for both `cli` and `jupyter` modes.
+- A helper script to convert LD matrices from old format to new format.
+
 ## [0.1.1] - 2024-04-12
 
 ### Changed

Diff for: containers/cli.Dockerfile

@@ -0,0 +1,43 @@
+# Usage:
+# ** Step 1 ** Build the docker image:
+# docker build -f cli.Dockerfile -t magenpy-cli .
+# ** Step 2** Run the docker container in interactive shell mode:
+# docker run -it magenpy-cli /bin/bash
+# ** Step 3** Test magenpy_ld:
+# magenpy_ld -h
+
+FROM python:3.11-slim-buster
+
+LABEL authors="Shadi Zabad"
+LABEL version="0.1"
+LABEL description="Docker image containing all requirements to run the commandline scripts in the magenpy package"
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    unzip \
+    wget \
+    pkg-config \
+    g++ gcc \
+    libopenblas-dev \
+    libomp-dev
+
+# Download and setup plink2:
+RUN mkdir -p /software && \
+    wget https://s3.amazonaws.com/plink2-assets/alpha5/plink2_linux_avx2_20240105.zip -O /software/plink2.zip && \
+    unzip /software/plink2.zip -d /software && \
+    rm /software/plink2.zip
+
+# Download and setup plink1.9:
+RUN mkdir -p /software && \
+    wget https://s3.amazonaws.com/plink1-assets/plink_linux_x86_64_20231211.zip -O /software/plink.zip && \
+    unzip /software/plink.zip -d /software && \
+    rm /software/plink.zip
+
+# Add plink1.9 and plink2 to PATH:
+RUN echo 'export PATH=$PATH:/software' >> ~/.bashrc
+
+# Install magenpy package from PyPI
+RUN pip install --upgrade pip magenpy
+
+# Test the installation
+RUN magenpy_ld -h

Diff for: containers/jupyter.Dockerfile

@@ -0,0 +1,53 @@
+# Usage:
+# ** Step 1 ** Build the docker image:
+# docker build -f ../vemPRS/containers/jupyter.Dockerfile -t magenpy-jupyter .
+# ** Step 2 ** Run the docker container (pass the appropriate port):
+# docker run -p 8888:8888 magenpy-jupyter
+# ** Step 3 ** Open the link in your browser:
+# http://localhost:8888
+
+
+FROM python:3.11-slim-buster
+
+LABEL authors="Shadi Zabad"
+LABEL version="0.1"
+LABEL description="Docker image containing all requirements to run the magenpy package in a Jupyter Notebook"
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    unzip \
+    wget \
+    pkg-config \
+    g++ gcc \
+    libopenblas-dev \
+    libomp-dev
+
+# Download and setup plink2:
+RUN mkdir -p /software && \
+    wget https://s3.amazonaws.com/plink2-assets/alpha5/plink2_linux_avx2_20240105.zip -O /software/plink2.zip && \
+    unzip /software/plink2.zip -d /software && \
+    rm /software/plink2.zip
+
+# Download and setup plink1.9:
+RUN mkdir -p /software && \
+    wget https://s3.amazonaws.com/plink1-assets/plink_linux_x86_64_20231211.zip -O /software/plink.zip && \
+    unzip /software/plink.zip -d /software && \
+    rm /software/plink.zip
+
+# Add plink1.9 and plink2 to PATH:
+RUN echo 'export PATH=$PATH:/software' >> ~/.bashrc
+
+# Install magenpy package from PyPI
+RUN pip install --upgrade pip magenpy jupyterlab
+
+# Expose the port Jupyter Lab will be served on
+EXPOSE 8888
+
+# Set the working directory
+WORKDIR /magenpy_dir
+
+# Copy the current directory contents into the container at /app
+COPY . /magenpy_dir
+
+# Run Jupyter Lab
+CMD ["jupyter", "lab", "--ip=0.0.0.0", "--allow-root", "--NotebookApp.token=''"]

Diff for: docs/installation.md

@@ -58,3 +58,21 @@ source magenpy_env/bin/activate
 python -m pip install --upgrade pip
 python -m pip install magenpy>=0.1
 ```
+
+### Using `Docker` containers
+
+If you are using `Docker` containers, you can build a container with the `viprs` package 
+and all its dependencies by downloading the relevant `Dockerfile` from the 
+[repository](https://github.com/shz9/magenpy/tree/master/containers) and building it 
+as follows:
+
+```bash
+# Build the docker image:
+docker build -f cli.Dockerfile -t magenpy-cli .
+# Run the container in interactive mode:
+docker run -it magenpy-cli /bin/bash
+# Test that the package installed successfully:
+magenpy_ld -h
+```
+
+We plan to publish pre-built `Docker` images on `DockerHub` in the future.

Diff for: examples/convert_old_ld_matrices.py

@@ -0,0 +1,69 @@
+"""
+This is a utility script that converts the old-style published LD matrices (magenpy 0.0.X) to the new
+format deployed since magenpy>=0.1. The old LD matrix format used ragged Zarr arrays, while the new format
+uses flattened Zarr arrays that are more efficient and easier to work with. The script takes the path to the
+old LD matrices and converts them to the new format with the desired precision (e.g. float32).
+
+The user may also specify the compressor name and compression level for the new LD matrices.
+The script will validate the conversion by checking the integrity of the new LD matrices.
+
+Usage:
+
+    python convert_old_ld_matrices.py --old-matrix-path /path/to/old/ld_matrices/chr_* \
+                                      --new-path /path/to/new/ld_matrices/ \
+                                      --dtype float32
+
+"""
+
+import magenpy as mgp
+from magenpy.utils.system_utils import makedir
+import zarr
+import os.path as osp
+import glob
+import argparse
+
+
+parser = argparse.ArgumentParser(description="""
+    Convert old-style LD matrices (magenpy 0.0.X) to the new format (magenpy >=0.1).
+""")
+
+parser.add_argument('--old-matrix-path', dest='old_path', type=str, required=True,
+                    help='The path to the old LD matrix. Can be a wild card of the form "path/to/chr_*"')
+parser.add_argument('--new-path', dest='new_path', type=str, required=True,
+                    help='The path where to store the new LD matrix.')
+parser.add_argument('--dtype', dest='dtype', type=str, default='int16',
+                    choices={'int8', 'int16', 'float32', 'float64'},
+                    help='The desired data type for the entries of the new LD matrix.')
+parser.add_argument('--compressor', dest='compressor', type=str, default='zstd',
+                    help='The compressor name for the new LD matrix.')
+parser.add_argument('--compression-level', dest='compression_level', type=int, default=9,
+                    help='The compression level for the new LD matrix.')
+
+args = parser.parse_args()
+
+for f in glob.glob(args.old_path):
+
+    try:
+        z_arr = zarr.open(f, 'r')
+        chrom = z_arr.attrs['Chromosome']
+    except Exception as e:
+        print(f"Error: {e}")
+        continue
+
+    print(f"> Converting LD matrix for chromosome: {chrom}")
+
+    new_path_suffix = f'chr_{chrom}'
+    if new_path_suffix not in args.new_path:
+        new_path = osp.join(args.new_path, new_path_suffix)
+    else:
+        new_path = args.new_path
+
+    makedir(new_path)
+
+    ld_mat = mgp.LDMatrix.from_ragged_zarr_matrix(f,
+                                                  new_path,
+                                                  overwrite=True,
+                                                  dtype=args.dtype,
+                                                  compressor_name=args.compressor,
+                                                  compression_level=args.compression_level)
+    print("Valid conversion:", ld_mat.validate_ld_matrix())

Diff for: magenpy/LDMatrix.py

@@ -2,6 +2,7 @@
 import os.path as osp
 import numpy as np
 import pandas as pd
+import warnings
 from scipy.sparse import csr_matrix, identity, triu, diags
 from .utils.model_utils import quantize, dequantize
 
@@ -712,27 +713,32 @@ def window_size(self):
     @property
     def n_neighbors(self):
         """
-        The number of variants in the LD window for each SNP.
-
         !!! seealso "See Also"
             * [window_size][magenpy.LDMatrix.LDMatrix.window_size]
 
         !!! note
             This includes the variant itself if the matrix is in memory and is symmetric.
 
+        :return: The number of variants in the LD window for each SNP.
+
         """
         return self.window_size()
 
     @property
     def csr_matrix(self):
         """
-        :return: The in-memory CSR matrix object.
-
         ..note ::
             If the LD matrix is not in-memory, then it'll be loaded using default settings.
+            This means that the matrix will be loaded as upper-triangular matrix with
+            default data type. To customize the loading, call the `.load(...)` method before
+            accessing the CSR matrix in this way.
 
+        :return: The in-memory CSR matrix object.
         """
         if self._mat is None:
+            warnings.warn("> Warning: Loading LD matrix with default settings. "
+                          "To customize, call the `.load(...)` method before invoking `.csr_matrix`.",
+                          stacklevel=2)
             self.load()
         return self._mat
 
@@ -833,7 +839,20 @@ def set_mask(self, mask):
         if self.in_memory:
             self.load(force_reload=True,
                       return_symmetric=self.is_symmetric,
-                      fill_diag=self.is_symmetric)
+                      fill_diag=self.is_symmetric,
+                      dtype=self.dtype)
+
+    def reset_mask(self):
+        """
+        Reset the mask to its default value (None).
+        """
+        self._mask = None
+
+        if self.in_memory:
+            self.load(force_reload=True,
+                      return_symmetric=self.is_symmetric,
+                      fill_diag=self.is_symmetric,
+                      dtype=self.dtype)
 
     def to_snp_table(self, col_subset=None):
         """
@@ -1409,11 +1428,11 @@ def validate_ld_matrix(self):
         return True
 
     def __getstate__(self):
-        return self.store.path, self.in_memory, self.is_symmetric, self._mask
+        return self.store.path, self.in_memory, self.is_symmetric, self._mask, self.dtype
 
     def __setstate__(self, state):
 
-        path, in_mem, is_symmetric, mask = state
+        path, in_mem, is_symmetric, mask, dtype = state
 
         self._zg = zarr.open_group(path, mode='r')
         self.in_memory = in_mem
@@ -1426,7 +1445,7 @@ def __setstate__(self, state):
             self.set_mask(mask)
 
         if in_mem:
-            self.load(return_symmetric=is_symmetric, fill_diag=is_symmetric)
+            self.load(return_symmetric=is_symmetric, fill_diag=is_symmetric, dtype=dtype)
 
     def __len__(self):
         return self.n_snps

Diff for: magenpy/SumstatsTable.py

@@ -310,7 +310,7 @@ def p_value(self):
         return self.pval
 
     @property
-    def log10_p_value(self):
+    def negative_log10_p_value(self):
         """
         :return: The negative log10 of the p-value (-log10(p_value)) of association
         test of each variant on the phenotype.
@@ -623,7 +623,7 @@ def to_table(self, col_subset=None):
             elif col == 'PVAL':
                 table['PVAL'] = self.p_value
             elif col == 'LOG10_PVAL':
-                table['LOG10_PVAL'] = self.log10_p_value
+                table['NLOG10_PVAL'] = self.negative_log10_p_value
             elif col == 'CHISQ':
                 table['CHISQ'] = self.get_chisq_statistic()
             elif col == 'MAF_VAR':

Diff for: magenpy/__init__.py

@@ -16,7 +16,7 @@
 
 from .utils.data_utils import *
 
-__version__ = '0.1.1'
+__version__ = '0.1.2'
 __release_date__ = 'April 2024'