The files in BED format can provide additional annotation for the sequence maps. 

Here is an example of the BED file with all 12 columns:


browser position chr1:253742-303353
browser hide all
track name="ItemRGBDemo" description="Item RGB demonstration" visibility=2 itemRgb="On"
chr1        253742        255091        teal_gene        0        +        253742        253742        250,200,100        2        104,155,        0,1194,
chr1        276321        303353        dblue_gene         0        +        284187        299335        0,0,200        8        73,203,148,137,129,156,184,4257,        0,5160,7845,12411,14326,15177,16713,22775,
chr1        276323        291537        cyan_gene        0        +        276323        276323        0,210,250        8        73,203,148,137,129,156,184,4257,        0,5160,7845,12411,14326,15177,16713,20775,
chr1        276352        291629        nothing        0        +        284187        291629        0        5        42,148,137,129,131,        0,7814,12380,14295,15146,

A minimum set of columns is: map_name, feature_start, feature_end

1        1000000        9990000
1        10000000        19990000
2        1001        2000
1        2000        3000

Colors

Note that the bed files allow creating annotation tracks with features that have individual colors (the eighth column). 

chr1        253742        255091        teal_gene        0        +        253742        253742        250,200,100
chr1        276321        303353        dblue_gene         0        +        284187        299335        0,0,200
chr1        276323        291537        cyan_gene        0        +        276323        276323        0,210,250
chr1        276352        291629        nothing        0        +        284187        291629        0

Unlike the marker tracks created in a similar way from the .csv files and also providing coloring capabilities, the bed file records can define multi-exon genes. 

thickStart and thickEnd

In case when the bed file is added as an external file, the CDS information ("thickStart" and "thickEnd") is ignored - just to avoid misinterpreting internal regions (e.g., providing protein sequences) which could mean something completely different from the CDS.

chr1        253742        255091        teal_gene        0        +        253742        253742        250,200,100
chr1        276321        303353        dblue_gene         0        +        284187        299335        0,0,200
chr1        276323        291537        cyan_gene        0        +        276323        276323        0,210,250
chr1        276352        291629        nothing        0        +        284187        291629        0