Look into bedtools window command. See https://bedtools.readthedocs.io/en/latest/content/tools/window.html. Should be pretty straight forward from there.

You could use bedmap from BEDOPS. Say you have a BED file of IS/transposable elements called elements.bed and a BED file of genes called genes.bed:

$ bedmap --echo --range 1000 --delim '\t' elements.bed genes.bed > answer.bed

The file answer.bed will have, on each line, a IS/transposable element and any genes which overlap a window 1kb upstream and downstream of the element's edges.

If you need genes in BED format and you are starting with annotations in GFF or GTF format, you can use gff2bed or gtf2bed, e.g.:

$ wget http://ftp.ensembl.org/pub/release-106/gtf/homo_sapiens/Homo_sapiens.GRCh38.106.chr.gtf.gz
$ gtf2bed < <(zcat Homo_sapiens.GRCh38.106.chr.gtf.gz) > Homo_sapiens.GRCh38.106.chr.bed

Or to avoid generating an intermediate file:

$ wget -qO- http://ftp.ensembl.org/pub/release-106/gtf/homo_sapiens/Homo_sapiens.GRCh38.106.chr.gtf.gz | gunzip -c | gtf2bed - > Homo_sapiens.GRCh38.106.chr.bed