1 / 19

ImageHTS Analysis of microscopy images of perturbed cell populations

ImageHTS Analysis of microscopy images of perturbed cell populations. ImageHTS. R package Analysis of microscopy images of perturbed cell populations Derived from Oleg's analysis framework Now Generic, works with any screen Extends cellHTS S4 object

bailey
Download Presentation

ImageHTS Analysis of microscopy images of perturbed cell populations

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ImageHTSAnalysis of microscopy images of perturbed cell populations

  2. ImageHTS • R package • Analysis of microscopy images of perturbed cell populations • Derived from Oleg's analysis framework • Now • Generic, works with any screen • Extends cellHTS S4 object • Can display and select cells using Remy's cellPicker • Emphasis on web-based distributed tools • To be released within 2 months

  3. Input data • Microscopy images of cells • Perturbed cells (siRNA, drugs) • CellHTS setup, images organized in (plate, sample, well) CD3EAP

  4. Pipeline • Classic pipeline • Segmentation of cells • Extraction of cell features • Cell classification • Summarization of cell features into population phenotypic profiles • Normalization • Reporting result (cellHTS report, quality score, hit list)

  5. Analysis script library(imageHTS) path = 'screens/remorpho' x = parseImageConf('imageconf.txt', path=path) x = configure(x, 'description.txt', 'plateconf.txt', 'screenlog.txt', path=path) unames = getUnames(x) ## segment wells, extract features and summarize features segmentWells(x, unames, 'segmentationpar.txt') extractFeatures(x, unames, 'summarizepar.txt') ## learning readLearnTS(x, 'trainingset.txt') ## predict cell labels predictCellLabels(x, unames) profiles = summarizeFeatures(x, unames, 'summarizepar.txt')

  6. Analysis configuration • Reusing cellHTS configuration files • plateconf.txt (screen geometry) • screenlog.txt (experimenter's screen log) • annotation.txt (reagent - target mapping) • Each module uses a configuration file seg.method: ath nuc.athresh.filter: makeBrush(35, shape='box')/(35*35) nuc.athresh.t: 0.00424 nuc.morpho.kernel: makeBrush(3, shape='diamond') nuc.watershed.tolerance: 3 nuc.watershed.neighbourood: 2 nuc.min.density: 0.1 nuc.min.size: 150.0625 nuc.max.size: 2070.25 ...

  7. File management • Images and intermediate data are huge (~1 TB) • Each project can be associated with a master URL where source images and intermediate results are stored • If not available locally, imageHTS will get files using this URL • Extremely convenient to analyze data from a remote computer • Several accessing modes (default, local, server, cache) neg = getUnames(x, type='neg') pos = getUnames(x, type='plk1') xneg = collectCellFeatures(x, neg, access='cache') yneg = collectCellFeatures(x, pos, access='cache') /home/gpau/projects/screens/remorpho/ behemoth2 user http://www.ebi.ac.uk/~gpau/private/aP123GregP/imageHTS/screens/remorpho/

  8. Extra tool: displayHTS • Display wells in a web page • Use cases: checking controls, tracking neg = getUnames(x, type='pos', plate=1:2) displayHTS(x, neg) profiles = loadHTS(x, 'profiles') displayHTS(x, profiles[,order(profiles$n.int)[1:10]])

  9. Extra tool: cellPicker • From Remy • Web-based client too to display/selecting cell • Use cases: tracking outlying cells, building a training set for classification • CellPicker is a distributable URL neg = getUnames(x, type='neg', plate=1:2) popCellPicker(x, neg, plabel='X')

  10. CellHTS report

  11. Query website

  12. Ongoing screens • Morphology screen • Hit retests on HeLa, U2OS cells, with replicates • Kinase screen on HeLa cells • Maybe MitocheckODE ? • DKFZ (Xian Zhang) • Kinase screens on mesenchymal cells • Kinase screen on stem cells

  13. Ongoing work + • Easy-to-use, fast, distributed framework • Now the software framework is done… • Interesting problems • Quality score of controls • Features selection • Features transformation • Hits • Distances between images • With sparse conditions, e.g. n = 300 wells, p = 200 features in a multi-parametric fashion feature 2 - feature 1

More Related