Learn

Talks & Tutorials

Pedagogical talks, tutorials, and infographics from BIA Lab and collaborators on computational methods, multiplexed imaging, quantitative cell biology, and practical tools like Claude Code.

Infographics

8 infographics
3D CyCIF infographic: see more markers, see more biology, in 3D
Clarence Yapp · Sorger Lab, HMS

3D CyCIF: see more markers, see more biology, in 3D

Laboratory of Systems Pharmacology, Harvard Medical School

A one-page visual summary of 3D CyCIF, which images many markers (20–50+) in the same intact 3D sample by repeating stain–image–erase cycles — preserving spatial context to map cell states, neighborhoods, and signaling niches in their true 3D architecture.

Paper (Yapp et al., 2025) Paper (Wong et al., 2026)
3D CyCIF multiplexed imaging spatial proteomics tumour microenvironment
Claude Code for Scientific Research infographic: a cheat sheet for using an AI coding agent in the lab
BIA Lab

Claude Code for Scientific Research

BIA Lab · Getting-Started Guide

A one-page cheat sheet for using an AI coding agent to speed up analysis, scripting, and data wrangling in the lab — covering the basic workflow, Plan Mode, usage limits, model tiers, and best practices for optimizing token usage.

Cheat Sheet (PDF)
View Download .docx
Slides (PDF)
View Download .pptx
Claude Code getting started computational workflows
SAM-SPOT infographic: a universal imaging phenome for dynamic cells
BIA Lab

SAM-SPOT: a universal imaging phenome for dynamic cells

BIA Lab · Software Infographic

A one-page visual summary of SAM-SPOT, the standardised framework for fingerprinting cell state at every moment using three complementary feature families — Shape, Appearance, and Motion.

Paper (SPOT) Paper (Organoids) PyPI GitHub
SAM-SPOT phenomics cell segmentation unsupervised clustering
MOSES infographic: motion sensing superpixels for quantifying collective cellular motion
BIA Lab

MOSES: motion sensing superpixels

BIA Lab · Software Infographic

A one-page visual summary of MOSES, which tracks dense optical flow through superpixel grids to build long-time motion trajectories and dynamic meshes — quantifying collective cellular motion phenotypes in time-lapse video without requiring single-cell segmentation.

Paper (eLife) Protocol PyPI GitHub
MOSES collective cell motion optical flow
u-Segment3D infographic: from 2D views to one accurate 3D segmentation
BIA Lab

u-segment3D: from 2D views to one accurate 3D segmentation

BIA Lab · Software Infographic

A one-page visual summary of u-segment3D, which combines 2D segmentations from one or more orthogonal views into a single consensus 3D cell labeling — compatible with any 2D segmentation method and requiring no 3D training.

Paper PyPI GitHub
u-segment3D 3D cell segmentation zero deep learning
u-Protrude3D infographic: automatic 3D protrusion segmentation and labeling from a triangulated cell surface mesh
BIA Lab

u-Protrude3D: automatic 3D protrusion segmentation & labeling

BIA Lab · Software Infographic

A one-page visual summary of u-Protrude3D, which automatically detects and labels individual protrusions — blebs, filopodia, microvilli, lamellipodia — on any triangulated 3D cell surface mesh, with no manual tracing required.

PyPI GitHub
u-Protrude3D protrusion segmentation 3D cell surface mesh
u-Unwrap3D infographic: see your cells in a whole new way
BIA Lab

u-Unwrap3D: see your cells in a whole new way

BIA Lab · Software Infographic

A one-page visual summary of u-Unwrap3D, which transforms complicated 3D cell surfaces into canonical representations — topographic maps, spherical parameterisations, and unfolded 2D images — for quantitative morphology and membrane-signal analysis.

Paper PyPI GitHub
u-Unwrap3D 3D surface parameterisation cell morphology
u-infotrace infographic: multiscale pixel spatiotemporal information flows
BIA Lab

u-infotrace: multiscale pixel spatiotemporal information flows

BIA Lab · Software Infographic

A one-page visual summary of u-infotrace, which estimates directional information flow between pixels across space and time — revealing causal influence patterns, information highways, and sources/sinks that optical flow alone cannot capture, via four interchangeable causal measures across a multiscale pyramid.

Paper GitHub (v-InfoTrace)
u-infotrace causal inference information flow spatiotemporal

Pipeline Talks

2 talks
Felix Y. Zhou · BIA Lab

Computational frameworks to quantitatively interrogate causal relationships between cell morphology and signaling

BioImaging North America · Lattice Light-Sheet User Group

Cells change shape. Signals move through them. The question this talk addresses is whether those two things are connected causally, and if so, in which direction. Felix presents the computational pipeline the lab has built to answer this: mapping molecular signals onto 3D cell surfaces at single-cell resolution, parameterising those surfaces into canonical coordinate systems to enable comparison across cells and time points, and applying information-theoretic measures to identify where and when morphology predicts downstream signaling. The talk draws on lattice light-sheet microscopy data and introduces tools including u-segment3D for universal 3D cell segmentation to extract a high-resolution cell surface, u-Unwrap3D for mapping signals onto those surfaces, and u-infotrace for quantifying information flow between signals.

3D surface parameterisation cell morphology information theory light-sheet microscopy causal inference
Clarence Yapp · Sorger Lab, HMS LifeCanvas Technologies

Decoding disease with highly-multiplexed 3D imaging

Laboratory of Systems Pharmacology, Harvard Medical School · LifeCanvas Technologies

Standard histology sees a handful of markers. This talk covers what becomes possible when you can measure one hundred or more proteins in intact tissue at subcellular resolution. Clarence Yapp from the Sorger Lab at Harvard Medical School presents the experimental and computational pipeline behind cyclic immunofluorescence (CyCIF) and volumetric multiplexed imaging: how tissue sections are repeatedly stained and imaged, how u-segment3D is used to identify every single cell in the volumetric tissue, how the resulting image stacks are registered and segmented, and how the spatial proteomics maps that emerge reveal the cell states, immune niches, and signaling architectures that distinguish diseased tissue from healthy tissue. The talk is a practical guide to the full stack, from antibody panel design to neighbourhood analysis.

CyCIF multiplexed imaging spatial proteomics tissue atlases tumour microenvironment

Tutorial Videos

1 tutorial
BIA Lab

SAM-SPOT: end-to-end tutorial from installation to phenotype analysis

BIA Lab · Software Tutorial Series

A practical walkthrough of the full SAM-SPOT pipeline. The tutorial covers installation, loading time-lapse imaging data, running automated cell segmentation, computing SAM (Shape, Appearance, Motion) feature representations for each cell, performing unsupervised clustering to identify phenotype groups, and visualising and comparing phenotype distributions across experimental conditions. Intended for researchers who want to get the pipeline running on their own data from scratch.

SAM-SPOT cell segmentation phenotype analysis time-lapse imaging unsupervised clustering

Visual Explainers

1 explainer
BIA Lab

SAM-SPOT: from raw imaging to phenotype landscapes

BIA Lab · Pipeline Schematic

An animated schematic of the full SAM-SPOT pipeline. Starting from the observation that cell state has three complementary axes — shape, appearance, and motion — the explainer walks through how 2185-dimensional SAM feature vectors are computed for every cell, how UMAP embeds all cells from all timepoints and timelapses into a single phenomic landscape, and how that landscape is then split into two analytical paths: discrete phenotype clustering with stacked-bar dynamics and Markov transition analysis, and continuous trajectory inference. The final scene shows how grouping SAM features into interpretable modules reveals which aspects of cell biology drive each phenotype cluster.

SAM-SPOT phenomics UMAP cell dynamics Markov transitions