Coverage by intersecting smaller genomic interval data over larger genomic intervals using R -

i want intersect 2 genomic intervals in r. , want coverage stats of smaller interval on larger interval.

the larger interval data data frame ....

chr  start     end       name         val    strand chr7 145444998 146102295 ccds5889.1   0      + chr7 146102406 146167735 ccds5889.1   0      + chr7 146167929 146371931 ccds5889.1   0      + 

the smaller interval more 2 million rows .

chr  start     end       name         val    strand phylop    chr7 145444386 145444387 ccds5889.1   0      +      0.684764 chr7 145444387 145444388 ccds5889.1   0      +      0.684764 chr7 145444388 145444389 ccds5889.1   0      +      0.684764 chr7 145444389 145444390 ccds5889.1   0      +      0.684764 

the interval data in 2nd (from) , 3rd (to) columns in both data frame.

the situation similar to

large interval:    [-----]   [-----]     [--------------]   [-------------------] small interval: |||  ||||  |||||||||||  ||||||||   ||||  || |||||||||   ||    |||||||| 

1. i want know how of each of larger intervals covered smaller intervals.
2. also, associate intersecting $phylop values each of large intervals later access plotting.

the bioconductor iranges , genomicranges packages have findoverlaps, countoverlaps, coverage , other interval-based functions designed these operations. you'd use granges function represent each of subject ("larger intervals") , query ("smaller intervals") objects above. see installation instructions , vignettes, e.g., browsevignettes("genomicranges")

in little more detail, here's data

sdf <- read.table(textconnection( "chr  start     end       name         val    strand chr7 145444998 146102295 ccds5889.1   0      + chr7 146102406 146167735 ccds5889.1   0      + chr7 146167929 146371931 ccds5889.1   0      +"), header=true)  qdf <- read.table(textconnection( "chr  start     end       name         val    strand phylop    chr7 145444386 145444387 ccds5889.1   0      +      0.684764 chr7 145444387 145444388 ccds5889.1   0      +      0.684764 chr7 145444388 145444389 ccds5889.1   0      +      0.684764 chr7 145444389 145444390 ccds5889.1   0      +      0.684764"), header=true) 

and here convert these granges , find intersection between query , subject

library(genomicranges) subj <-     with(sdf, granges(chr, iranges(start, end), strand, val=val)) query <-     with(qdf, granges(chr, iranges(start, end), strand, val=val,                       phylop=phylop, names=name)) intersect(query, subj) 

the answer isn't exciting here

> intersect(query, subj) granges 0 ranges , 0 elementmetadata values      seqnames ranges strand |  seqlengths  chr7    na 

to little more useful, here's query tiles across total region

tiles <- successiveiranges(rep(100, 950), 900, 145444998) query <- granges("chr7", tiles, "+") 

we find intersecting ranges, find subject overlapped each intersecting range, , summarize width of overlapping ranges

int <- intersect(query, subj) tapply(int, subjecthits(findoverlaps(int, subj)),        function(x) sum(width(x))) 

leading to

    1     2     3  65800  6500 20400