English to Russian Science Mathematics & Statistics

Step 1:
Independent component analysis (ICA)
The ICA found an ‘unmixing’ matrix W that linearly unmixed the input data matrix X (7176 x 117K) into a spatially independent source matrix U, where U=WX.
The rows of estimated source matrix U (component activations) were the waveforms of the corresponding ICs along each of the markers.
At this step, the ICA analysis returned 7176 components for further analysis.
In ICA, the component processes are assumed to sum to the measured methylation values, without pre-specified noise terms, though some components may in fact include or represent one or more types of ‘noise’ in the data.
Step 2:
Z-transform standardization to the component activation
In order to fairly assess the contribution of each marker among 7176 ICs, the Z-transform standardization to the component activations was applied.
Specifically, each component activation (one row of U) removed its mean and divided the value by the standard deviation to have zero mean and unit variance.
This procedure generated a normalized component activation U (i.e., marker weightings) in the so-called Z-values.

Как я понял, это разновидность элемента матрицы.