Matt:LabNotes/2015-3-12

From ZhangLabWiki
Jump to navigation Jump to search

Dextran Sulfate in dcProbe Hybridization to Improve Signal Uniformity (bimodal distribution)[edit]

  • Many FISH protocols (eg Lubeck, Cai 2014) use 10% Dextran Sulfate in hybridization buffer
    • Theory is dextran sulfate excludes H2O thus increasing effective concentration of probes near the tissue
  • Hope is that adding dextran sulfate will give a bimodal distribution of pixel intensities making it easy to determine a threshold for calling "OFF" states

Experiment[edit]

  • Sample: BA8_S2_V4_3-4-2015
  • Dye-probe: 0.5uM FISGA_Adpt (Cy3)
    • 3 Conditions:
      • Dye 1) 0.5uM in 30% formamide + 2X SSC (1ul 100uM dye + 20ul 20X SSC + 60ul formamide + 119ul H2O = 200ul)
      • Dye 2) 0.5uM in 10% dextran sulfate + 30% formamide + 2X SSC (0.5ul 100uM dye + 10ul 20X SSC + 30ul formamide + 20ul 50% dextran sulfate + 39.5ul H2O = 100ul)
      • Dye 3) 0.5uM in 10% dextran sulfate + 10% formamide + 2X SSC (0.5ul 100uM dye + 10ul 20X SSC + 10ul formamide + 20ul 50% dextran sulfate + 59.5ul H2O = 100ul)

Imaging Order[edit]

  1. Strip sample with 80% formamide
  2. Add 100ul Dye 1 to find 4 positions and set laser power & gain to get minimal saturated pixels while broadest distribution of intensities
    • 1% laser (552nm wavelength), [V] = 590, % offset = -1.00
  3. Image (Named Step 1)
  4. Strip
  5. Add 100ul Dye 1 and Image (Named Step 2)
  6. Strip
  7. Add 100ul Dye 2 and Image (Named Step 3)
  8. Strip
  9. Add 100ul Dye 3 and Image (Named Step 4)
  10. Strip
  • For all:
    • preheat dyes & 80% formamide in 2X SSC to 75C
    • incubate @RT 10min to hybridize and then wash with 2X SSC twice
    • incubate @RT 15min to strip and then wash with 1X PBS twice
  • Take 10 z-stack images for each of 4 positions (40 images for each step)

Results[edit]

  • Saved images in folder 3-12-2015
  • Aggregate all 40 images for each step and plot histogram ( Matlab script)

Log scale Y axis (intensity)[edit]

  • No significant differences

File:DextranExp Step1 LogYHist.JPGFile:DextranExp Step2 LogYHist.JPG
File:DextranExp Step3 LogYHist.JPGFile:DextranExp Step4 LogYHist.JPG

Y axis zoomed in to show details of higher intensity counts[edit]

  • No significant differences
    • Step 4 (10% formamide 10% dextran sulfate) has higher intensity pixels than Step 3 (30% formamide 10% dextran sulfate)

File:DextranExp Step1 ZoomYHist.JPGFile:DextranExp Step2 ZoomYHist.JPG
File:DextranExp Step3 ZoomYHist.JPGFile:DextranExp Step4 ZoomYHist.JPG

Ideas/Thoughts[edit]

  • Deconvolution could improve distribution
  • Try w/ PISA mask to find threshold (distribution of PISA identified features vs distribution of background)
  • Shorten RCA time to decrease max value will also decrease variance
  • Try subtracting background (image after stripping or with non-matching dcProbes as background)
    • Use streptavidin fluorescent beads fixed with BS(PEG)9 to use as landmarks for alignment