| abstract |
The invention discloses a rapid and low-loss group single-cell big data reduction method, belonging to the technical field of single-cell big data reduction, comprising the following steps: S1, cell subgroup screening; S2, data dimensionality reduction; S3, cell subgroup screening Subpopulation reduction intensity value i setting; S4, cell cluster construction; S5, cell data reduction. With this reduction method, the amount of single-cell data in a population can be reduced based on cell clusters on the premise of minimal loss of cellular heterogeneity information within a subpopulation, thereby improving the efficiency of downstream single-cell data analysis. |