It's like "computeMatrix" in deeptools, computeMatrix scale-regions -R TAD.boundaries.bed -S gene.density.bw gives the distribution of the eigenvalues (in this case genes) represented by the -S parameter over the genomic positions (in this case TAD boundaries) represented by the -R parameter. (Then I get the distribution of genes on the TAD boundary). But deeptools doesn't run fast and takes up a lot of memory.

I have tried using TAD.boundaries.bed instead of peak.bed and ChIPseeker actually runs much faster than deeptools and takes up very little memory compared to (So I get the distribution of the TAD boundary upstream and downstream of the TSS).

I then tried to manually change the TAD.boundaries.bed to gff format as input to "makeTxDbFromGFF" and then extracted the location of each gene as peak.bed as input to "readPeakFile", but it gives me an error: "在为'viewSums'函数选择方法时评估'x'参数出了错: subscript is out of bounds".

If this were possible, more functions could be implemented to map the distribution of any feature value of the genome at any position in the genome.