Question

featAssayPlotDefaults does not work for SummarizedExperiment object

1

Entering edit mode

rishi.dasroy ▴ 20

@rishidasroy-7142

Last seen 7 months ago

Finland

I was trying to follow the vignett for customizing iSEE interface and then I encountered following error with reproducible code.

Construct a SummarizedExperiment object

nrows <- 200
ncols <- 6
counts <- matrix(runif(nrows * ncols, 1, 1e4), nrows)
se <- SummarizedExperiment(assays=list(counts=counts))

Now try a function from ISEE package meant for SummarizedExperiment objects

fexArg <- featAssayPlotDefaults(se, 1)

which produces following error

Error in (function (classes, fdef, mtable) : unable to find an inherited method for function 'int_metadata' for signature '"SummarizedExperiment"'

How to fix this? I am using iSEE version 1.6.1

Many thanks

iSEE SummarizedExperiment • 1.5k views

ADD COMMENT • link updated 4.5 years ago by Aaron Lun ★ 28k • written 4.5 years ago by rishi.dasroy ▴ 20

score 3 · Accepted Answer · 2020-04-24

3

Entering edit mode

Aaron Lun ★ 28k

@alun

Last seen 5 hours ago

The city by the bay

Short answer: just coerce your SE to a SingleCellExperiment instead.

Long answer: the next release of iSEE (2.0.0) provides a much smoother way of setting up the app:

iSEE(se, initial=list(
     FeatureAssayPlot(),
     ColumnDataPlot(),
     RowDataTable()
))

These functions will replace the old *Defaults(), and should avoid running into the error above.

ADD COMMENT • link 4.5 years ago Aaron Lun ★ 28k

0

Entering edit mode

On the contrary, the documentation of featAssayPlotDefaults says it will process SummarizedExperiment, that is the reason for my confusion.

ADD REPLY • link 4.5 years ago rishi.dasroy ▴ 20

0

Entering edit mode

To clarify:

SummarizedExperiment are indeed the core class supported by iSEE(), which means that derived classes such as SingleCellExperiment are implicitly supported.

As a derived class, the SingleCellExperiment class inherits from all the features of the parent SummarizedExperiment class, while adding new slots and methods specific to the SingleCellExperiment. One such new method is int_metadata(). In other words, int_metadata() is defined for the derived SingleCellExperiment class, but not for the parent SummarizedExperiment.

To conclude, Aaron's advice to coerce your SE to a SingleCellExperiment will add SingleCellExperiment-specific slots to your object, which will then enable the associated methods, including int_metadata().

Please follow that advice. We'll have a closer look at the documentation, to see if we can make it clearer for new users.

ADD REPLY • link 4.5 years ago kevin.rue ▴ 350

0

Entering edit mode

Thanks, Aron and Kevin for replying. However, with the below code, I have tried to coerced to SingleCellExperiment which have not produced "Feature assay plot 1" .

library(airway)
data(airway)

airway_sce <- SingleCellExperiment(airway)
iSEE::iSEE(airway_sce)

The warnings in the console are

Listening on http://127.0.0.1:5613 'Reduced dimension plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Column data plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Feature assay plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Row data plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Sample assay plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Custom data plot 1' in 'initialPanels' is not available (maximum ID is 0) 'Custom statistics table 1' in 'initialPanels' is not available (maximum ID is 0) 'Heat map 1' in 'initialPanels' is not available (maximum ID is 0)

ADD REPLY • link 4.5 years ago rishi.dasroy ▴ 20

1

Entering edit mode

The correct command is

airway_sce <- as(airway, "SingleCellExperiment")

In versions ≤1.6.1, iSEE requires assays to be named, which is your issue in this instance.

Please use R 4.0.0 with Bioconductor release 3.11 and iSEE 2.0.0, available today, which brings a new simpler syntax and new features.

ADD REPLY • link 4.5 years ago kevin.rue ▴ 350

0

Entering edit mode

It has worked finally. Many thanks to both of you.

ADD REPLY • link 4.5 years ago rishi.dasroy ▴ 20

0

Entering edit mode

Good news! Can I please ask you to 'accept' this answer to mark this question as resolved? Thank you !

ADD REPLY • link 4.5 years ago kevin.rue ▴ 350