Tracking endosomes in hippocampal neurons Yannis Kalaidzidis - PowerPoint PPT Presentation

Tracking endosomes in hippocampal neurons Yannis Kalaidzidis 2019-01-09 QBI-2019 Rennes

Major challenge - low SNR

Major challenge - low SNR Livia Goto-Silva

Major challenge - low SN

Intensity time course 180 175 Intensity (a.u.) 170 165 160 155 150 t (sec) 0.0 20 40 60 80 100

Frog Eye Filter: Probabilistic model   I F B     2 2    I F b F   I b  1  1 1            2  2 , | , , 2 2 P I b e e e    2 2

Hyperparameter estimation  2           1  1 1 2 1  2 t t   t      2         2     erfc   erfc   1 2 t e e t e         2 2   2 2  m t  2            1 1 t t    2     2 erfc    1  erfc     e e t          2 2 2 from equation we found β and μ .

Background estimation   2     1 I F at b   1              2 2 | , , , , | , P I a b e P F dF I F B  2   0 B at b i   1      R       1 2 1  2 e d   R       | , , , , erfc   2 P I a b e     2 2 2    at b I       where ; ; R d R       1     R       i i i           2 N 1 1 e d                 1 i   R d log erfc   1 L d     i i i   i     N  t 2  L  2      i i 1 i 0 i        1   a    1 1  i d i  i  1  2 x i  2 2 e       where   x  1          1  1  R d   erfc i i i   x   N 1 L       i 2 0      1   b    1 i 1 d i   i  i

Background estimation 1200 1000 800 600 400 200 0.0 5.0 10 15 20 25

Estimation of foreground intensity                  , , | , , , , , | , , , | , | , P I F B B I P I F B I P F P B B 0 0 0 0 b b                  | , , , , , , , | , , , , , P F b I P I F B b I dB dI 0 0 b b 0 0      2     1 I B F      F 2 2  0 0   I F B 1    2   2 2    0 0 b erfc , 0 e e F b                   2 2 2 2 2 2 2      b b b       | , , , , , P F B I 0 0 b      2    I B   1     0 0 2 2 1   I B     2   2    2 b 0 0 erfc , 0 e F F b                  2 2 2 2 2 2 2    b b b

Estimation of foreground intensity           | , , , , , F F P F B I dF 0 0 b 0    2 1    2 W 1     2    2 2 1 erfc 1 e  T      2   2        I B s G        0 0   1 1       2 1 erfc   erfc   G U       2 2    F   1      1 G       s I B s I              2 2 2 0 0 where ; ; ; ; s G T     b 2 1 s s b       2   1 B s 2 I B  0  0 ; W U   s s

De-noised movie

De-noising Evaluation 220 Raw Images 200 180 160 140 120 0.0 1.0 2.0 3.0 4.0 220 200 Frog Eye Filter 180 160 140 120 0.0 1.0 2.0 3.0 4.0

De-noising Evaluation 220 Raw Images 200 180 160 140 120 0.0 1.0 2.0 3.0 4.0 220 200 PURE-LET 180 160 140 120 0.0 1.0 2.0 3.0 4.0

De-noising Evaluation 220 200 PURE-LET 180 160 140 120 0.0 1.0 2.0 3.0 4.0 220 200 Frog Eye Filter 180 160 140 120 0.0 1.0 2.0 3.0 4.0

De-noising Evaluation CTV PURE-LET Raw Frog Eye (ICY) (Fiji)

Intensity time course in ROI Original Image 185 180 Intensity (a.u.) 175 170 165 160 155 150 0.0 20 40 60 80 100 PURE-LET Frog Eye time (sec) Endosomes cross through ROI 185 185 180 180 Intensity (a.u.) Intensity (a.u.) 175 175 170 170 165 165 160 160 155 155 150 150 0.0 20 40 60 80 100 0.0 20 40 60 80 100 time (sec) time (sec)

High background challenge

High background challenge

Frog Eye Filter 175 170 Intensity (a.u.) 165 160 155 66 68 70 72 74 76 78 80 82 t (sec)

Frog Eye Filter 175 170 Intensity (a.u.) 165 160 155 66 68 70 72 74 76 78 80 82 t (sec)

Frog Eye Filter 175 170 Intensity (a.u.) 165 160 155 66 68 70 72 74 76 78 80 82 t (sec)

Frog Eye Filter

Frog Eye Filter

Frog Eye Filter

Frog Eye filter (APPL1) retrograde retrograde

Speed distribution Retrograde Anterograde 2000   0.358 0.055 m r 1500   # movements 0.331 0.056 m a  Student_t 0.73 p value 1000 500 0.0 0.01 0.1 1.0 Speed (µm/sec)

Speed distribution 2000 1500 # movements 1000 500 0.0 0.01 0.1 1.0 Speed (µm/sec)

Speed distribution Retrograde µ δµ σ δσ A δ A m δ m 0.936 0.052 0.69 0.035 7 494 543 1.18 0.072 0.087 0.003 1.388 0.023 45 583 554 0.22 0.01 2000 1500 # movements 1000 500 0.0 0.01 0.1 1.0 Speed (µm/sec)

Speed distribution Retrograde µ δµ σ δσ A δ A m δ m 0.936 0.052 0.69 0.035 7 494 543 1.18 0.072 0.087 0.003 1.388 0.023 45 583 554 0.22 0.01 2000 Anterograde µ δµ σ δσ A δ A m δ m 1.674 0.099 0.44 0.043 2 881 392 1.84 0.11 1500 # movements 0.085 0.002 1.42 0.027 43 662 596 0.24 0.01 1000 500 0.0 0.01 0.1 1.0 Speed (µm/sec)

Speed distribution Retrograde µ δµ σ δσ A δ A m δ m 0.936 0.052 0.69 0.035 7 494 543 1.18 0.072 0.087 0.003 1.388 0.023 45 583 554 0.22 0.01 2000 Anterograde µ δµ σ δσ A δ A m δ m 1.674 0.099 0.44 0.043 2 881 392 1.84 0.11 1500 # movements 0.085 0.002 1.42 0.027 43 662 596 0.24 0.01 1000    7 Fast Movement: Student_t 9.6 10 p value  Slow Movement: Student_t 0.157 p value 500 0.0 0.01 0.1 1.0 Speed (µm/sec)

Conclusion • De-noising allow to compensate low SNR and image in low phototoxicity/low bleaching mode. • Frog Eye filter allow track dim endosomes in axons with high background fluorescence • Sped distribution de-convolusion allows to reveal significantly different components, which are not distinguishable in average values

Acknowledgements Marino Zerial Lab Hernán Andrés Morales Navarrete Alexander Kalaidzidis http://motintracking.mpi-cbg.de