MOTCO

Multi-omics Trajectory Comparison — a Python package for building latent spaces from multi-omics data and quantifying group differences in those spaces.

What it does

MOTCO provides three statistical modules:

Module	Purpose
PLS-DA	Build supervised latent spaces via double cross-validated PLS discriminant analysis
SNF	Fuse multiple omics layers into a single similarity network
Trajectory Analysis	Estimate and test differences in trajectory magnitude, orientation, and shape between groups
Orientation Driver Attribution	Explain shared-PLS trajectory orientation differences with signed feature, module, and pathway effects
Simulations	Invoke R InterSIM and return aligned semi-synthetic omics matrices

Quick links

• GitHub repository
• End-to-end example notebook
• CLI reference

Installation

pip install motco
# or with uv:
uv add motco

API Reference

• PLS-DA — plsda_doubleCV, calculate_vips
• SNF — get_affinity_matrix, SNF, get_spectral
• Trajectory Analysis — get_model_matrix, build_ls_means, estimate_difference, RRPP, estimate_betas, get_observed_vectors, pair_difference, center_matrix
• Orientation Driver Attribution — analyze_orientation_attribution, attribution_frames, write_attribution_outputs
• Simulations — InterSIMParams, InterSIMResult, check_intersim_available, run_intersim
• CLI — motco plsr, motco snf, motco de