Function to extract spatial summary functions from the Spatial slot of an mxFDA object
Arguments
- mxFDAobject
object of class
mxFDA- r_vec
Numeric vector of radii over which to evaluate spatial summary functions. Must begin at 0.
- extract_func
Defaults to univariate, which calculates univariate spatial summary functions. Choose bivariate for bivariate spatial summary functions.
- summary_func
Spatial summary function to calculate. Options are c(Kest, Lest, Gest) which denote Ripley's K, Besag's L, and nearest neighbor G function, respectively.
- markvar
The name of the variable that denotes cell type(s) of interest. Character.
- mark1
Character string that denotes first cell type of interest.
- mark2
Character string that denotes second cell type of interest for calculating bivariate summary statistics. Not used when calculating univariate statistics.
- edge_correction
Character string that denotes the edge correction method for spatial summary function. For Kest and Lest choose one of c("border", "isotropic", "Ripley", "translate", "none"). For Gest choose one of c("rs", "km", "han")
- empirical_CSR
logical to indicate whether to use the permutations to identify the sample-specific complete spatial randomness (CSR) estimation. If there are not enough levels present in
markvarcolumn for permutations, the theoretical will be used.- permutations
integer for the number of permtuations to use if empirical_CSR is
TRUEand exact CSR not calculable
Value
an object of class mxFDA containing the corresponding spatial summary function slot filled. See make_mxfda() for object structure details.
Details
![[Stable]](figures/lifecycle-stable.svg)
Complete spatial randomness (CSR) is the estimation or measure of a spatial summary function when the points or cells in a sample are randomly distributed, following no clustering or dispersion pattern. Some samples do have artifacts that may influence what CSR is under the distribution of points as they are found in the sample such as large regions of missing points or possibly in the case of tissue sections, necrotic tissue where cells are dead. Theoretical CSR requires points have an equal chance of occurring anywhere in the sample that these artifacts violate, necessitating the need to estimate or calculate what this CSR would be for each sample independently. Previously Wilson et al. had demonstrated cases in which sample-specific CSR was important over the use of the theoretical in calculating how much the observed deviates from expected.
References
Xiao, L., Ruppert, D., Zipunnikov, V., and Crainiceanu, C. (2016). Fast covariance estimation for high-dimensional functional data. Statistics and Computing, 26, 409-421. DOI: 10.1007/s11222-014-9485-x.
Wilson, C., Soupir, A. C., Thapa, R., Creed, J., Nguyen, J., Segura, C. M., Gerke, T., Schildkraut, J. M., Peres, L. C., & Fridley, B. L. (2022). Tumor immune cell clustering and its association with survival in African American women with ovarian cancer. PLoS computational biology, 18(3), e1009900. https://doi.org/10.1371/journal.pcbi.1009900
Creed, J. H., Wilson, C. M., Soupir, A. C., Colin-Leitzinger, C. M., Kimmel, G. J., Ospina, O. E., Chakiryan, N. H., Markowitz, J., Peres, L. C., Coghill, A., & Fridley, B. L. (2021). spatialTIME and iTIME: R package and Shiny application for visualization and analysis of immunofluorescence data. Bioinformatics (Oxford, England), 37(23), 4584–4586. https://doi.org/10.1093/bioinformatics/btab757
Examples
#load ovarian FDA object
data('ovarian_FDA')
#run function
ovarian_FDA = extract_summary_functions(ovarian_FDA, r_vec = 0:100,
extract_func = univariate,
summary_func = Gest,
markvar = "immune",
mark1 = "immune",
edge_correction = "rs")
#> Using Theoretical Complete Spatial Randomness for Nearest Neighbor G