Bioc 2023- Spot Deconvolution

Slides covering spot deconvolution benchmarking and implementation, as presented at the Bioconductor 2023 conference.

Transcript

1. Spot Deconvolution in the Post-Mortem Human DLPFC Nicholas J. Eagles Research Associate nick-eagles.github.io @Nick-Eagles 1 https://speakerdeck.com/nickeagles
2. 2 Study Design • Neurotypical Adults • High quality RNA (RIN > 7) • 30 Visium samples (10 donors x 3 positions) • 19 snRNA-seq samples (10 donors x 2 positions) • 4 Visium-SPG samples (separate donors) -SPG (SRT + IF)
3. Integrating SRT and Single-Cell Data to Map Cell Types Spatially - Single-cell data lacks spatial information, and SRT data lacks cell-type-composition information - Spot deconvolution determines cell composition within Visium spots single-cell spatial Questions we can’t answer from SRT data alone: How are cellular populations distributed spatially? Which cell types are communicating in ligand-receptor interactions associated with schizophrenia? Image from Bo Xia: https://twitter.com/BoXia7/status/1261464021322137600?s=12 3 spot deconvolution
4. Spot Deconvolution: Visually 4 Br6522_Ant_IF
5. Sean Maden @MadenSean Sang Ho Kwon @sanghokwon17 #deconvochallenge doi.org/10.48550/arXiv.2305.06501 5
6. Benchmark existing spot deconvolution methods, and choose the best to estimate cell-type composition in our spatial data 6 My Goal
7. Benchmarking Spot Deconvolution Software: Theory - How do we measure performance or accuracy of cell-type predictions? - Make orthogonal measurements*: image-derived counts - Leverage prior knowledge: neurons localize to gray matter 7 Stephanie C. Hicks Benchmarked Software Tangram (Biancalani et al.) Cell2location (Kleshchevnikov et al.) SPOTlight (Elosua-Bayes et al.)
8. Visium-SPG IF Images as an Orthogonal Measurement 8
9. Using IF to Quantify Cell Types - Visium-SPG IF images mark for several proteins - Fluorescence in image channels correlates with counts of measured cell types Can measure 5 distinct cell types: 9 • Astrocyte (GFAP) • Neuron (NeuN) • Oligodendrocyte (OLIG2) • Microglia (TMEM119) • Other (low signal in all channels) samuibrowser.com Chaichontat Sriworarat Stephanie C. Hicks doi.org/10.1101/2023.01.28.525943
10. Segmenting Cells on Visium-SPG IF Images 1. Segment cells on IF image 2. Manually label example cells 3. Train cell-type classifier and apply on remaining data 10
11. Constructing Dataset of Labeled Cells 11 1. Segment cells on IF image 2. Manually label example cells 3. Train cell-type classifier and apply on remaining data 4 sections * 5 cell types * 46 cells = 920 manually labeled cells Annie B. Nguyen
12. Training Cell-Type Classifier 12 1. Segment cells on IF image 2. Manually label N cells 3. Train cell-type classifier and apply on remaining data Final model chosen
13. Benchmark Results 13
14. Benchmark Results: Make Orthogonal Measurements 14 Micro (Br6522_Ant_IF) All points in A -> one point in B
15. Benchmark Results: Leverage Prior Knowledge - Manually annotate spots with histological layer - Explore how cell-type predictions map to annotated layers Kristen R. Maynard Br6522_Ant_IF 15
16. 16 Max across layers Not max Benchmark Results: Leverage Prior Knowledge
17. Benchmark Results: Leverage Prior Knowledge 17
18. Benchmark Summary 18 Metric Tangram Cell2location SPOTlight Metric Type Avg. spot-level cor. 0.31 0.30 0.21 Orthogonal measurements Avg. spot-level RMSE 1.35 1.24 1.3 Orthogonal measurements Histological mapping 0.69 0.77 0.23 Leverage known biology Tangram Cell2location SPOTlight
19. Data Availability Samui Browser Shiny Apps https://libd.shinyapps.io/spatialDLPFC_Visium_SPG/ 19
20. Conclusions - Imaging data can be leveraged to infer cell-type composition in Visium data - Existing spot deconvolution algorithms have limited accuracy - Need for benchmarks and benchmarking best practices Check out our preprint! 20 2:00 pm Package Demo Analyzing Spatially-Resolved Transcriptomics Data from Visium using spatialLIBD Louise A. Huuki-Myers
21. Acknowledgements LIBD Annie B. Nguyen Leonardo Collado-Torres Kristen R. Maynard Louise Huuki-Myers Abby Spangler Kelsey D. Montgomery Sang Ho Kwon Heena R. Divecha Madhavi Tippani Matthew N. Tran Arta Seyedian Thomas M. Hyde Joel E. Kleinman Stephanie C. Page Keri Martinowich JHU Biostatistics Chaichontat Swirorarat Stephanie C. Hicks Boyi Guo JHU Biomedical Engineering Alexis Battle Prashanthi Ravichandran PsychENCODE consortium University College London Genetics and Genomic Medicine Mina Ryten Melissa Grant-Peters nick-eagles.github.io @Nick-Eagles 21 https://speakerdeck.com/nickeagles
22. Model Test Precision Test Recall Decision tree 0.86 0.87 Logistic regression 0.91 0.90 Support vector machine 0.90 0.90 Grid search with 5-fold CV for each model to select hyperparameters Model # Training # Test Split 720 200 ~78/22 Data More about dataset and model