Data set

[bethac07]

Raw images: https://open.quiltdata.com/b/cellpainting-gallery/tree/cpg0031-caicedo-cmvip/broad/images/BBBC043/images/
Segmentation masks: are per site, one example: https://open.quiltdata.com/b/cellpainting-gallery/tree/cpg0031-caicedo-cmvip/broad/workspace/analysis/BBBC043/52652/analysis/52652-a01-4/outlines/

These commands will get you row A from any plate, feel free to switch rows (the lowercase letter in the include commands; you can do any lowercase letter a through p), and the plate number listed above (52652) is as far as I know as good as any (but you might ie want one/some plates for train and some for test). You'll need to do some extra file moving to get all the masks in one folder but shouldn't be onerous.:
To get the segmentation masks for nuclei:
aws s3 sync s3://cellpainting-gallery/cpg0031-caicedo-cmvip/broad/workspace/analysis/BBBC043/{PLATE}/analysis/ . --exclude="*" --include="*-a*-*nuclei_label_matrix.tiff"
To get the raw images:
aws s3 sync s3://cellpainting-gallery/cpg0031-caicedo-cmvip/broad/images/BBBC043/images/{PLATE}/ . --exclude="*" --include="*_a*_w1*" --exclude="*thumb*"