... | ... | @@ -13,9 +13,9 @@ Each image file must contain an image of at most 3 dimensions with the third dim |
|
|
▷ well002
|
|
|
▷ plate1_replicate2
|
|
|
|
|
|
then the image root directory is screen_images.
|
|
|
then the image root directory is screen_images. The organization of images under the root directory can follow any arbitrary structure as long as the
|
|
|
|
|
|
The IDE can in principle read all image formats supported by BioFormats but has so far only been tested with TIFF.
|
|
|
The IDE can in principle read all image formats supported by BioFormats but has so far mostly been tested with TIFF.
|
|
|
|
|
|
The image root directory must be accessible from the computer running the app.
|
|
|
|
... | ... | @@ -23,15 +23,17 @@ The image root directory must be accessible from the computer running the app. |
|
|
##### Format
|
|
|
Image-derived data are expected to be in table format with data points in rows and stored in a **tab- or comma-separated text file using ASCII or UTF-8 encoding**. The table must have **column headers with unique column names** and all columns must have a header. For numbers, the decimal separator must be . (dot) and no separator for thousands is allowed.
|
|
|
If the data is distributed across multiple tables in the same directory, these can be combined into one table with the following shell commands run in the directory containing the tables:
|
|
|
`(head -1 fileA.csv && tail -n +2 -q file*.csv) >> ../table.csv` or
|
|
|
`awk 'NR==1 || FNR!=1' *.csv > ../table.csv` or
|
|
|
`(head -1 fileA.csv && tail -n +2 -q file*.csv) >> ../table.csv`
|
|
|
or
|
|
|
`awk 'NR==1 || FNR!=1' *.csv > ../table.csv`
|
|
|
or
|
|
|
for doing it recursively (i.e. when the tables are in multiple subdirectories):
|
|
|
`find . -type f -name "*.csv" -exec awk 'NR==1 || FNR!=1' {} + > ../table.csv`
|
|
|
Note that the tables must have the same format, i.e. use the same delimiter and have the same columns in the same order. This also assumes that the tables are the only .csv files in the directory.
|
|
|
##### Content
|
|
|
To link data points to images, the table should include one column with the **path to image files relative to the image root directory** and each cell of this column must reference only one file. Using the example above, the image root directory is 'screen\_images' and therefore the table column for data points associated with image W001-P001-Z000-T0000-s1234-Cy3.tif should contain the relative path 'plate1_replicate1/well001/W001-P001-Z000-T0000-s1234-Cy3.tif'. There can be multiple columns with links to images but only two can be used simultaneously in the IDE.
|
|
|
The IDE references ROIs by the **coordinates of an anchor point** (e.g. the ROI centre) therefore there should be a column for each of the relevant coordinates: x, y and either z or t. Coordinates (x,y) must be in pixels relative to the top left corner of the image (which is pretty much the standard for image analysis software).
|
|
|
For high-throughput microscopy data, the table should have separate columns for plate and well identifiers. Wells must be identified either by indices starting from 1 in the top left corner of the plate and incremented by row from left to right or by row/column coordinates with rows identified by letters and columns by numbers with well A1 in the top left corner of the plate. All wells are expected to be represented in the data file.
|
|
|
For high-throughput microscopy data, the table should have separate columns for plate and well identifiers. Wells must be identified either by indices starting from 1 in the top left corner of the plate and incremented by row from left to right or by row/column coordinates with rows identified by letters and columns by numbers with well A1 in the top left corner of the plate. **All wells must be represented in the data file. Missing wells may cause unexpected behaviour.**
|
|
|
|
|
|
|
|
|
|