Nov 2020 Discussions (associations with PDCD11)

[jatinarora-upmc]

@shntnu @AnneCarpenter @raldanehme @bethac07

I would like to seek your guidance on one more association (hopefully the last one or among the lasts). Rare variant burden in PDCD11 gene is associated with higher granularity from mito channel in cell and nuclei body in Isolate cells.

PDCD11 is required for rRNA maturation and generation of 18S rRNA.

Q1: in the microscopy images, i have pointed isolate cells with white arrows. The (isolate) cells with disrupted PDCD11 are brighter than wild-type cells, and you can not even notice nucleus as well as in wild-type cells. Do you think this brightness can proxy granularity?

[jatinarora-upmc]

Q1: i guess i would need to compare bright-field images of ER channel separately.

[shntnu]

Do you think this brightness can proxy granularity?

Granularity is intensity invariant (i.e. if you double the intensity, the granularity metric won't change. See https://forum.image.sc/t/cellprofiler-granularity-measurements/36966/2 for details.)

There is however a relationship between the standard deviation of intensity and granularity, which you can see empirically (below). But you are talking about overall brightness (i.e. the mean or sum), and granularity is independent of that.

df <- read_csv("1.profile-cell-lines/data/cmQTLplate7-2-27-20_sampled.csv.gz")

df1 <- df %>%
  select(matches("Cells_Granularity_[0-9]+_Mito|Cells_Intensity_.*_Mito"))

corrplot::corrplot(cor(as.matrix(df1)), type = "full", tl.pos = "l")

[AnneCarpenter]

Those metrics of granularity of the mito channel in th nucleus/cell indeed are indicating that the mitochondria staining pattern is more textured for the variant. It fits what you see because in the WT case the stain is excluded from the nucleus and so the intensity levels are likely very smooth. I cannot elaborate on what this "means" to have the mitochondrial staining textured and spread throughout the cell like that, personally.