Hosuk:LabNotes/2014-8-19

From ZhangLabWiki
Revision as of 22:24, 22 August 2014 by >Hosuki78
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search


Decoding with Agi26k0gap Padlock Probe[edit]

  • Image resolution : 20x obj, 4096 x 4086, z step = 1.04um
    • Since how many features would be detected or how much dense rolonies in this sample, so image in wide field of view
    • eventually 63x obj.
  • Data is in '2014-08-13'


Analysis[edit]

  • 1. Arrange data
    1. File name (decided at imaging) : [Sample]_[Resolution]_[Step#]_[Pos#]_z##_ch##.tif
      • ex) PGP1F : PGP1F_Agi26k0gap_Decode_P20_S2_2014-08-09_20x_4k_Step1_Pos2_z01_ch01.tif
      • ex) Tissue section : HBTissue_2014-08-12_RCAagain_20x_4k_Pos1_z00_ch00.tif
    2. Step# after confocal imaging
      • Step0 : 1st Rolony --> ch00 : Cy3, ch01 : BF
      • Step1~7 : ch00 : 488, ch01 : Cy3, ch02 : Cy5, ch03 : BF
    3. Maximum Intensity Projection(MIP) at each channel
      • Re-format the Step1~Step7 --> Step1~21
      • Step1 --> Step1:488, Step2:Cy3, Step3:Cy5
      • Step2 --> Step4:488, Step5:Cy3, Step6:Cy5
      • ...Step7 --> Step19:488, Step20:Cy3, Step21:Cy5
    4. Image aligning
      • Step# in one cycle is an exact position, no shift happened --> Cycle N = Step[3(N-1)+1], Step[3(N-1)+2], Step[3(N-1)+3]
      • Get offset using Bright field images (ch03) of each Cycle : Step1, Step4, Step7, ... Step19
        • I've done by manually, but eventually need automation
    5. Shift, Crop of ch00, ch01, ch02 images(Image_Aligning_Crop_v5.m)
      • Name : Crop_[file name]
      • ex)Offset
Cycle X Y
1 0 0
2 -13 +8
3 -9 +8
4 -19 +4
5 -16 +4
5 -15 +10
5 -6 -9
Average Offset -11.14 +0.71
Round -11 +1
New Offset
1 -11 +1
2 -2 +7
3 +2 +7
4 -8 +3
5 -5 +3
5 -4 -11
5 +5 -10
Crop Size 12 12


  • 2. Run Script for detected barcodes and counts(RolonyDataInfo_v12_7StageDecoding.m, Rolony_Decoding_v18c_7StageDecoding.m)
    1. Run PISA with every image (Step1~21)
      • PISA parameter
      • PISAParam.log_radius = 2;
      • PISAParam.log_upper = -2e-4;
      • PISAParam.area_upper = 100;
      • PISAParam.area_lower = 10;
      • PISAParam.axratio_lower = 0.5;
      • PISAParam.circ_upper = 1.6;
      • PISAParam.circ_lower = 0.8;
      • PISAParam.pconn = 8;
      • PISAParam.bkgmult_lower = 3;
      • PISAParam.FilterOption = 1;
      • PISAParam.Deconvoluted = 0;
      • PISAParam.DoMedflt = 1;
      • PISAParam.nofindbkg = 0;
      • eventually masking each image with MIP image will be used
    2. Result gemerated : Barcodes, Counts


  • 3. Mapping detected barcode string to gene
    1. Matt made master key table : barcode and gene name
    2. Allow 1 bit error (--> but these barcodes of 1 bit error were found in other genes...)
    3. Map detected barcode to gene name using VLOOKUP(E#, Sheet0Ex!$A$2:$B$24000, 2, 0)


Result[edit]

  • Saved at 2014-08-18
    • Total num : 2471
    • Num detected only one dye at each digit : 1100
    • Num detected 488 and Cy3 at each digit : 775
    • Num detected 488 and Cy5 at each digit : 785
    • Num detected Cy3 and Cy5 at each digit : 275
    • Num detected 488, Cy3 and Cy5 at each digit : 99
  • Mapped to gene : 348
  • Need to consider features with two or three dyes to be mapped to gene --> need more to polish the data


File:DecodeResultSummary 2014-08-18.png