Merging adjacent BEDGRAPH intervals with coverage >= mincov

daniel.sh

# assumes that the BEDGRAPH is the output of the bedtools genomecov tool using the -bga option.
$ cat foo.bedg
chr1	0	10	5
chr1	10	20	5
chr1	20	30	10
chr1	30	40	11
chr1	40	50	9
chr1	50	60	10
chr1	60	70	11
chr1	70	80	15
chr1	80	90	5
chr1	90	100	10

# 1. add an extra column with the MINCOV (e.g., 10) if the depth meets MINCOV
$ cat foo.bedg | awk -v mincov=10 '{ if ($4 >= mincov) {print $0"\t"mincov} else {print $0"\t"$4}}'
chr1	0	10	5	5
chr1	10	20	5	5
chr1	20	30	10	10
chr1	30	40	11	10
chr1	40	50	9	9
chr1	50	60	10	10
chr1	60	70	11	10
chr1	70	80	15	10
chr1	80	90	5	5
chr1	90	100	10	10

# 2. we can now use the extra column as a grouping criterion (assumes BEDGRAPH)
# 3. group on the surrogate column and report a new interval with the min start and max end in the "block"
# 4. Yank out the relevant columns.
# 5. Profit.
$ cat foo.bedg \
     | awk -v mincov=10 '{ if ($4 >= mincov) {print $0"\t"mincov} else {print $0"\t"$4}}' \
     | bedtools groupby -g 1,5 -c 1,2,3,4 -o first,first,last,first \
     | cut -f 3-6
chr1	0	20	5
chr1	20	40	10
chr1	40	50	9
chr1	50	80	10
chr1	80	90	5
chr1	90	100	10

# You could also use mean as the new column.
$ cat foo.bedg \
     | awk -v mincov=10 '{ if ($4 >= mincov) {print $0"\t"mincov} else {print $0"\t"$4}}' \
     | bedtools groupby -g 1,5 -c 1,2,3,4 -o first,first,last,mean \
     | cut -f 3-6
chr1	0	20	5
chr1	20	40	10.5
chr1	40	50	9
chr1	50	80	12
chr1	80	90	5
chr1	90	100	10

$ cat foo.bedg \
     | awk -v mincov=10 '{ if ($4 >= mincov) {print $0"\t"mincov} else {print $0"\t"$4}}' \
     | bedtools groupby -g 1,5 -c 1,2,3,4 -o first,first,last,max \
     | cut -f 3-6
chr1	0	20	5
chr1	20	40	11
chr1	40	50	9
chr1	50	80	15
chr1	80	90	5
chr1	90	100	10