About how long were HGP DNA pieces before insertion into BACs?

Before insertion into bacterial artificial chromosomes, HGP DNA pieces were about 150,000 base pairs long.

The public HGP used a hierarchical, or clone-based, shotgun strategy. Researchers first broke the genome into large fragments, typically around 150 kilobases, and inserted those fragments into BACs. Each BAC could be propagated inside bacteria as a stable clone, providing a physical piece whose approximate chromosomal location could be mapped before detailed sequencing.

The BAC inserts were not the final sequencing reads. Scientists selected mapped BACs, sheared or subcloned their DNA into much smaller fragments, sequenced those fragments, and computationally assembled the overlapping results. The known order of BAC clones supplied a scaffold that made assembling a roughly three-billion-base-pair genome more manageable.

The figure is an average, not an exact size for every clone. A widely used human BAC library, RPCI-11, had an average insert of about 178 kilobases, while many HGP descriptions use approximately 150,000 base pairs as the representative fragment length. That distinction explains why both numbers appear in technical accounts.