Welcome to Morphops!

Morphops implements common operations and algorithms for Geometric Morphometrics, in Python 3.

Build status Documentation Status PyPI version

Features

Some high-level operations in the current version are

Dependencies

  • numpy

Installation

pip install morphops

Usage Examples

import morphops as mops

# Create 3 landmark sets, each having 5 landmarks in 2 dimensions.
A = [[0,0],[2,0],[2,2],[1,3],[0,2]]
B = [[0.1,-0.1],[2,0],[2.3,1.8],[1,3],[0.4,2]]
C = [[-0.1,-0.1],[2.1,0],[2,1.8],[0.9,3.1],[-0.4,2.1]]

# Perform Generalized Procrustes alignment to align A, B, C.
res = mops.gpa([A, B, C])
# res['aligned'] contains the aligned A, B, C. res['mean'] is their mean.

# Create a Thin-plate Spline warp from A to B and warp C.
warped_C = mops.tps_warp(A, B, C)
# warped_C contains the image of the pts in C under the TPS warp.

What is Geometric Morphometrics?

Geometric Morphometrics is a statistical toolkit for quantifying and studying shapes of forms that are represented by homologous landmark sets.

“Shape” has a specific notion here. For a given landmark set, its shape refers to the spatial information that survives after discarding its absolute position, scale and rotation. So two landmark sets have the same shape if they can be brought in perfect alignment by only changing their positions, scales and rotations.

Common Operations and Algorithms in Studies

Geometric Morphometrics is often used when pursuing statistical questions involving the morphology of biological forms, like do corvid species that frequently probe have longer bills and more to-the-side orbits than corvid species that frequently peck. It helps inform the Data Collection, Preprocessing and Analysis steps of such statistical studies with sound theoretical or practical justifications.

Data Collection

The most prevalent form of Data Collection involves picking homologous landmarks on each form. For curving forms with few homologous points but well-understood homologous regions, there is a notion of semilandmarks which can “slide” to minimize equidistant sampling artifacts.

A common file format for saving landmarks for a set of specimens is the *.dta format used by the IDAV Landmark Editor software.

Preprocessing

As discussed before, a central idea in Geometric Morphometrics is extracting the “shapes” of the landmark sets. One way to achieve this is to use the Generalized Procrustes Alignment algorithm or GPA. GPA aligns all the landmark sets by modifying their locations, orientations and sizes so as to minimize their collective interlandmark distances.

After this step, the aligned shapes all lie in a high-dimensional non-linear manifold. For example, if the orignal landmark sets were a set of triangles, the aligned shapes lie on a sphere. Moreover, for naturally arising datasets, the shapes likely lie very close to each other and are distributed around a mean shape. This usually makes it permissible to project all the shapes into the tangent space at the mean shape, and this way the final shape vectors lie in a linear space.

Analysis

With the shapes lying in a high-dimensional linear space after preprocessing, they can now be submitted to various commonly used statistical procedures like Principal Components Analysis and various kinds of regression for further analysis.

Contributors ✨

Thanks goes to these wonderful people (emoji key):


Vaibhav Patel

Håkon Wiik Ånes

📖 🔧 🚇 🚧

This project follows the all-contributors specification. Contributions of any kind are welcome!

This list is maintained manually until such time that the all-contributors bot supports rst. A possibly fix may be coming in PR 301 there.

morphops module

The morphops module is the primary module of the Morphops library.

It contains implementations of common geometric morphometrics operations. Some examples are -

  • IO operations to read/write landmark data
  • Common preprocessing like Generalized Procrustes Alignment
  • Thin-plate spline warping operations

Submodules

procrustes Provides procrustes alignment related operations and algorithms.
tps Provides thin-plate splines related operations and algorithms.
io Provides IO functions to read from and write to files in common landmark data file formats.
lmk_util Provides common functions used in the module.
morphops.VERSION = '0.1.13'

The version of this module.

Indices and tables