The examples compare empirical findings across multiple dimensions of a pre-specified 240-cell data matrix which consists of content, spatial and temporal dimensions.
In each data set, samples (S) are rows and species/taxa (SP) columns; each cell in the data matrix contains an abundance value (x, e.g. numbers of individuals, coverage, cell numbers or presence/absence).
= percentage of cells in the data matrix scored as missing or inapplicable.
These remaining biclusters cover only 4.7% of cells in the data matrix.
Compared to that, searched biclusters from BCCA only covered 77.6% of genes and 31.7% of cells in the data matrix although 100% of conditions were covered by biclusters.
Only highly up-regulated or down-regulated genes in limited conditions, which are about 11.2% of cells in a data matrix, are searched in QUBIC.
Although average size of biclusters in CPB was 8413.6, those biclusters only cover 51.2% of genes and 18.5% of cells in the data matrix.
The SNPs s are ordered by their physical positions along the chromosome, with the cells of the data matrix X s, i taking the value 0 if heterozygous, and arbitrarily −1 or 1 for the alternative homozygous states.
If each row is for a cell line in a data matrix, we have four columns (cell line rank per subtype) of the ranks as seen in Supplementary Table 14 and the ranks are plotted in Figure 3, shown per cell line.
Out of 339,561 data matrix cells (= character states), 6% were polymorphic and 30% were missing (unknown or inapplicable).
The transmit symbols and precoding matrices of all served users are combined in the cell data vector and the cell precoding matrix x i = x 1, i T, …, x S i, i T T ∈ ℂ ℓ i, (3) F i = F 1, i, …, F S i, i ∈ ℂ N i × ℓ i. (4).
