I have a dataset of >300 4-channel composites that I have built a pipeline to process. The pipeline works great and is able to identify the features I am concerned about. Unfortunately, when I went to apply the pipeline to my entire dataset I realized that some of the images were taken with a different channel-order in the SP5 software. Instead of the Ch0-Ch3 order being “B-G-R-W” (W=white), the channel order is “R-W-B-G.” The result is that my pipeline attempts to quantify nuclei (DAPI) from the red ( R) channel image and so on. The analysis essentially does not work on those parts of the dataset.
In short, the way that the pipeline works is that it counts objects based on “B” and then counts objects in “G” masked by “B,” to classify all cells expressing “G.” It then masks “B” and “G” using “R” and “W” to classify “R/W” cells with “B” and subsequently “G.” The channel assignments are generated using the “Names-Types” module and I am not currently using any metadata.
I’m wondering if there is a way to extract the channel information from a “.LIF” file or to pair the single-channel images with their associated LUT files to assign channel names in place of the number (with a color rather than Ch00 etc.) or if there is a way to extract that information from a metadata file (I don’t know of a way to generate a meta-.CSV file from LAS-X)?
I have tried using bio-formats to convert the metadata into a CSV file…but cellprofiler doesn’t collect the channel-names from that file…it rather just collects “monochrome” vs. “RGB.”
Images are currently stored as “Tiff” files, they were not exported as "imageJ Tiff"