Some of "K" plates the Z scores are really low

[tischi]

@metavibor @constantinpape @imagirom I am starting to look into it...

plate_name                        `median(IgG_robust_z_score_means, na.rm = T)`
  <chr>                                                                     <dbl>
1 20200417_132123_311                                                       0.968
2 20200417_152052_943                                                       0.980
3 20200417_203228_156                                                       0.936
4 plate6rep2_wp_20200507_131032_010                                         1.40 
5 plate9_2rep1_20200506_163349_413                                          2.75 
6 plate9rep1_20200430_144438_974                                            1.55 
7 plateK12rep1_20200430_155932_313                                          0.585
8 PlateK19rep1_20200506_095722_264                                          1.21

I can reproduce the low z-scores of plate plateK12rep1_20200430_155932_313

[tischi]

@imagirom @metavibor Can we somehow shorten the plate names? For plots like below it is quite annoying that they are so long...

[constantinpape]

We should not rename anything internally at this point, that will be a horror for data-managements. But we can come up with shorter handles for the plates to use in the manuscript.

[tischi]

In plate K12 the mad is relatively high:

# A tibble: 8 x 2
  plate_name                   `median(IgG_control_mad_means, na.rm = T)`
  <chr>                                                             <dbl>
1 20200417_132123_311                                                146.
2 20200417_152052_943                                                197.
3 20200417_203228_156                                                112.
4 plate6rep2_wp0507_131032_010                                       291.
5 plate9_2rep10506_163349_413                                        162.
6 plate9rep10430_144438_974                                          667.
7 plateK12rep10430_155932_313                                        526.
8 PlateK19rep10506_095722_264                                        268.

...while the difference between infected and control is relatively low:

# A tibble: 8 x 2
  plate_name                   `median(IgG_diff_means, na.rm = T)`
  <chr>                                                      <dbl>
1 20200417_132123_311                                         151.
2 20200417_152052_943                                         172.
3 20200417_203228_156                                         100.
4 plate6rep2_wp0507_131032_010                                412.
5 plate9_2rep10506_163349_413                                 466.
6 plate9rep10430_144438_974                                  1193.
7 plateK12rep10430_155932_313                                 266.
8 PlateK19rep10506_095722_264                                 326.

thus one should indeed check whether the infected cell classification worked in this plate

[tischi]

see here: https://github.com/hci-unihd/antibodies-analysis-issues/issues/66