Single-particle tracking · PALM

Track single molecules.
Trust the numbers.

Tracelux finds, links and analyses single fluorescent molecules in Zeiss .czi and TIFF movies — from raw frames to diffusion coefficients, diffusive states and statistics. On PALMTracer's own data, it finds the same spots and builds the same trajectories.

See how it compares Get in touch

macOS (Apple silicon) and Windows · Coming soon

The Tracelux start page: analyse a sample, batch a folder, compare and analyse, or visualise tracks.

Validated against PALMTracer

Matches PALMTracer, spot for spot.

Tracelux and PALMTracer analysed the same 16 recordings, with PALMTracer's own detection threshold for each file and the same linking limits: a 3 px search range, no gap closing, trajectories of at least 8 points. Nothing was tuned or calibrated on this data.

99.93%

of 3.37 million spots found at the identical position

99.2%

of 109,413 trajectories reproduced exactly

99.9%

of 97,041 shared trajectories with a diffusion coefficient within 1%

Left: the distribution of diffusion coefficients of every trajectory; Tracelux's curve lies on top of PALMTracer's. Right: trajectory-by-trajectory diffusion coefficients from both programs fall on the identity line.
Every trajectory from all 16 recordings (left), and each trajectory both programs found, Tracelux against PALMTracer (right). The few differences are where two molecules pass close together and the two linkers choose differently.

What it does

The whole analysis, in one app.

Detection that matches PALMTracer

A PALMTracer-style wavelet detector, plus Crocker–Grier, Gaussian maximum-likelihood and radial-symmetry engines. Uses the GPU on Apple silicon and NVIDIA cards where available.

Six linkers

Crocker–Grier, Kalman filter, Jaqaman LAP with or without merging and splitting, nearest-neighbour and simulated annealing.

Diffusion and motion

Per-trajectory MSD, diffusion coefficients and anomalous exponents, motion classes, jump-distance distributions, dwell times, MSS and velocity autocorrelation.

Diffusive states

A three-state hidden Markov model of every recording — immobile, slow and fast mobile — with each state's diffusion coefficient, occupancy and switching probabilities.

Live comparison

Drop in 2 to 12 conditions. Figures, effect sizes and p-values corrected for multiple comparisons update as you add data — no "generate" step.

Publication figures

A 19-panel figure for every recording, per-panel exports, comparison reports and colour-blind-safe palettes.

Drift and clustering

Redundant cross-correlation drift correction and DBSCAN clustering of localisations, with per-cluster area and density.

HYPER-LUX batch

On a large workstation a folder of recordings fans out across worker processes, each shown live on its own dashboard tile.

How it works

Four steps from movie to figure.

  1. 01

    Import

    Pick a recording or a folder. Pixel size and frame interval come from the file; preview detection and draw regions of interest.

  2. 02

    Process

    Detection, drift correction, linking, diffusion and classification run in the background, with a live view of each stage.

  3. 03

    Analysis

    Compare conditions in one live view: figures, statistics and a ready-to-copy methods sentence.

  4. 04

    Visualise

    Play back every trajectory coloured by motion class, explore tracks, clusters and super-resolution reconstructions.

Reproducibility

Every result knows where it came from.

Each analysis writes a manifest beside its results, so a figure can always be traced back — and re-made.

  • Exact version. The Tracelux version and the version of every library that can change a number.
  • Every setting. All detection, linking and analysis parameters, as used.
  • The input file. Its SHA-256 checksum, so you know it is the same recording.
  • Replay. Load a manifest to repeat a run exactly — including runs made by FIREFLY, Tracelux's former name.

Coming soon.

Tracelux is in private testing ahead of release for macOS (Apple silicon) and Windows. For early access, collaborations or questions, get in touch.

hello@tracelux.app