The raw file you downloaded,
opened with your own hands
The figures in a paper are results somebody else has already finished drawing. The real learning is upstream of them. Download GSE268555 from GEO and what you are actually holding is four .h5 files โ with 40,100 cells and 32,285 genes inside them. How do you open that, and what do you look at first?
These 20 lessons begin at the moment you open that one file. Python runs right inside this page.
How you'll learn
Not an imitation. 1,200 cells ร 32,285 genes, cut straight out of the original h5 deposited at GEO as GSE268555, opened by you with h5py.
What to check the moment a file lands in your hands โ groups, datasets, shapes, dtypes, compression. You get the order in which you read somebody else's file into your fingers.
A table of numbers tells you nothing. The sliding gallery widget lets you page through sixteen views of this dataset and see the whole of it in thirty seconds.
The moment memory blows up, the moment gene names collide, the moment a batch effect fools you โ you live through the accidents that really happen in advance.