EpiTraceAge_Convergence

Main workhorse function in the package.

Input:

  • peakSet: a prepared “peak set” GRanges object.

  • matrix: a prepared read count matrix.

  • celltype: optional annotation for cell types. Usually not required.

  • sep_string: used to split the peak names. Usually as c(“:”,”_”,”-“).

  • clock_gr_list: GRange list for reference clock-like loci. A list.

  • non_standard_clock: set to T when you need to use non-default clock-like loci as reference. NOTE the standard clock-like loci is only useful when you work on human data.

  • qualnum: minimal peaks that a cell/sample should have.

  • remove_peaks_number: minimal number of cells that a peak should be positive on.

  • ref_genome: when it is set to “hg19” or “hg38”, the program can automatically use the clock-like loci from package. Otherwise, you should provide your own reference clock-like loci.

  • run_reduction: set to F if you would not like Signac dimensionality reduction to run.

  • fn.k.param: KNN cluster parameter for Signac.

  • lsi_dim: LSI dimensions to be used in Signac dimensionality reduction.

  • min.cutoff: minimal cutoff for Signac::FindTopFeatures. Used when you ask the program to run dimensionality reduction.

  • Z_cutoff: cutoff during iteration to select putative new clock-like loci. Defined as: include all loci with a scaled correlation coefficient peak x age > Z_cutoff. Usually set to 2~3.

  • mean_error_limit: limit of mean sample age differences between iterations. If the difference is smaller than this, iteration stops early.

  • iterative_time: maximal iteration time.

  • normalization_method: choose between “randomized”, “census”, and “blank”. Default to “randomized”.

  • select_minimal_percentage: minimal cell fraction that showing coverage on the selected peak. Default to 0.05.

  • select_size_of_dispersion: number of peaks selected for computing the similarity matrix, ranked by dispersion.

Output: a Seurat object with ages annotated in its metadata.

Main informations added to the meta.data of the Seurat objects are:

  • EpiTraceAge_Clock_initial: the age estimated using only reference clock-like loci

  • EpiTraceAge_iterative: the age estimated using reference + newly-included (putative sample-specific) clock-like loci