Matrix File Format

FlatProt uses transformation matrices to control how protein structures are oriented in the 2D visualization. This document explains the format of these matrix files.

Overview

A transformation matrix in FlatProt consists of:

  • A 3×3 rotation matrix
  • A 3-element translation vector

Together, these define how the protein structure is positioned and oriented in 3D space before projection to 2D.

File Format

Matrix files should be saved as NumPy .npy files containing a 2D array with one of the following shapes:

  1. 4×3 matrix (preferred format):

    • First 3 rows: 3×3 rotation matrix
    • Last row: 1×3 translation vector

  2. 3×4 matrix (will be automatically transposed):

    • First 3 columns: 3×3 rotation matrix
    • Last column: 3×1 translation vector

  3. 3×3 matrix (rotation only):

    • The entire matrix is treated as a rotation matrix
    • Translation is assumed to be zero

Creating Matrix Files

You can create a matrix file using NumPy:

import numpy as np

# Create a rotation matrix (identity rotation)
rotation = np.eye(3)

# Create a translation vector (move 5 units along x-axis)
translation = np.array([5.0, 0.0, 0.0])

# Combine into a 4×3 matrix
matrix = np.vstack([rotation, translation])

# Save to a file
np.save("my_transformation.npy", matrix)

Validation

FlatProt validates matrix files to ensure they have the correct dimensions and format. If a matrix file is invalid, an error message will be displayed explaining the issue.