Skip to content

DicotPipeline

Pipeline for analyzing dicot plants with primary and lateral roots.

Overview

DicotPipeline is designed for dicot species (e.g., soy, canola, arabidopsis) that have a single primary root with branching lateral roots.

Root types required:

  • Primary root (1 instance)
  • Lateral roots (multiple instances)

Computed traits: 40+ morphological measurements

Quick Start

import sleap_roots as sr

# Load SLEAP predictions
series = sr.Series.load(
    series_name="919QDUH",
    h5_path="919QDUH.h5",
    primary_path="919QDUH.primary.slp",
    lateral_path="919QDUH.lateral.slp"
)

# Create pipeline and compute traits
pipeline = sr.DicotPipeline()
traits = pipeline.compute_plant_traits(series, write_csv=True)

print(traits[['frame_idx', 'primary_length', 'lateral_count']].head())

Trait Categories

Length Traits

  • primary_length – Total length of primary root
  • lateral_length_total – Sum of all lateral root lengths
  • lateral_length_avg – Average lateral root length
  • lateral_length_max – Maximum lateral root length
  • lateral_length_min – Minimum lateral root length

Count Traits

  • lateral_count – Number of lateral roots detected
  • primary_tip_count – Number of primary root tips

Angle Traits

  • primary_angle – Angle of primary root from vertical
  • lateral_angle_avg – Average emergence angle of laterals
  • lateral_angle_min – Minimum lateral emergence angle
  • lateral_angle_max – Maximum lateral emergence angle

Topology Traits

  • convex_hull_area – Area of convex hull around all roots
  • network_distribution_ratio – Spatial distribution metric
  • network_solidity – Compactness of root network

Data Requirements

SLEAP Skeleton

Primary root skeleton: 

base → node1 → node2 → ... → tip

Lateral root skeleton (multiple instances): 

base → node1 → node2 → ... → tip

File Organization

Single time-point: 

plant_name.primary.slp  # Primary root predictions
plant_name.lateral.slp  # Lateral root predictions
plant_name.h5           # Optional: video file

Example Datasets

The test data includes dicot examples:

  • Canola (7 days old): tests/data/canola_7do/
  • Soy (6 days old): tests/data/soy_6do/

Best Practices

  • Ensure lateral roots are tracked separately from primary
  • Label lateral root bases near primary root
  • Use consistent skeleton structure across plants
  • Verify SLEAP predictions before batch processing

See Also