The progress of the Human Genome Project was faster than expected (see Box1 for a timeline). Genetic maps were developed ahead of the original schedule, and the final stage of large-scale DNA sequencing was facilitated by developments in automated fluorescence-based DNA sequencing. Competition between publically and privately funded sequencing programs also drove a rapid final sequencing phase.
BOX1. MAJOR MILESTONES IN MAPPING AND SEQUENCING THE HUMAN GENOME
An important rationale of the HGP—and a major motivation for privately funded genome sequencing—was to be able to study human genes. Starting in the early 1990s, attempts were made to obtain partial sequences from the 3′ untranslated regions of as many different human cDNA clones as possible, generating a huge number of ESTs. Introns are rarely found in the 3′ untranslated region of human genes and so a PCR assay based on the EST sequence could usually be used to type genomic DNA. Subsequently, systematic large-scale mapping of ESTs against panels of radiation hybrids produced the first comprehensive human gene maps. The resulting gene map was published in 1998 (see Box1) and appeared to identify the positions of 30,000 human genes. However, the full extent of our genes could not be known with more precision until the genome sequence was delivered.
From an early stage it was clear that human genes were not uniformly distributed along or between chromosomes. Some chromosomes were rich in genes; others were gene-poor (Figure 1). The heterochromatic regions of the genome—including most of the Y chromosome, and substantial regions on chromosomes 1, 9, and 16—were known to be essentially devoid of genes and extraordinarily rich in repetitive DNA that would make mapping extremely difficult. As a result, the HGP was almost exclusively focused on the remaining euchromatic regions that collectively accounted for about 90% of the human genome.
Fig1. An early human gene map. Many human (and other vertebrate) genes have associated CpG islands (sequences about 1 kb long, often at the 5′ ends of genes, that differ from the bulk of the DNA in having many unmethylated CpG dinucleotides). The image shows the result of hybridizing a purified human CpG island fraction (labeled with a Texas Red stain) to human metaphase chromosomes. Late-replicating chromosomal regions (mostly transcriptionally inactive) are distinguished by incorporation of FITC labeled bromodeoxyuridine (green signal). Yellow regions (overlap of red and green signals) denote late-replicating regions rich in genes (or strictly, CpG islands). Because CpG islands are gene markers, chromosomal regions that show a strong red signal have a high gene density (e.g., chromosome 22). Other chromosomes have very weak red signals and are gene-poor, such as chromosomes 4, 18, X, and Y. (Adapted from Craig JM & Bickmore WA [1994] Nat Genet 7:376–382; PMID 7920655. With permission from Springer Nature. Copyright © 1994.)
For the publically funded International Human Genome Sequencing Consortium (IHGSC), most of the sequence was contributed by large genome centers. To ensure efficiency it was agreed that specific centers would take primary responsibility for the assembly of clone contigs and subsequent sequencing of individual chromosomes: for example, the Wellcome Trust Sanger Institute for chromosome 1, Washington University for chromosome 2, and so on.
A rival, commercial human genome sequencing effort was announced in 1999 when the Celera company declared its intention to produce a draft human genome sequence in 2 years: a whole-genome shotgun sequencing approach was to be used instead of the hierarchical shotgun sequencing approach of the IHGSC (which required BAC clones to be ordered into contigs before carrying out shotgun sequencing of the large inserts, which was more time-consuming). The ensuing race between the IHGSC and Celera accelerated the timetable. In 2001 both sides published a draft sequence of the human genome that covered about 90% of the euchromatic genome sequence. The euchromatic component is ~90% of the total genome, and so the draft sequences actually represented about 80% of the total genome, but 90% of the total gene sequence.
Although the race to achieve a draft genome sequence was perceived to have ended in a draw in 2001, it had not been a fair race, and the finishing line had not been reached— there was still some hard work to be done. The IHGSC had made their data freely available to all, posting sequence data updates on the Web every 24 hours. Celera took huge blocks of the IHGSC’s sequence data, reprocessed it, and fed the data back into its own sequence compilation. The Celera sequence was, therefore, not an independently obtained human genome sequence. Unlike the IHGSC, Celera denied free external access to their sequence data (and continued to require expensive subscription charges to view their sequence data long after they had published their analyses).
The hard work toward finishing the euchromatic human genome sequence was carried out by the IHGSC, leading to publication of the virtually complete sequence in 2004. Even by the end of 2018, however, the human euchromatic genome sequence remained unfinished, largely because of genome assembly problems. Some of the problems arose because the source DNA was not a haploid genome (which would have been ideal for assembly purposes), and not even a diploid sequence of an individual person (for ethical reasons). Instead, a variety of different people contributed blood-cell samples for making the large-insert libraries used to provide the final DNA templates for sequencing (Box 2). Different regions of the human genome sequence therefore originated from different individuals, and structural variation between haplotypes impeded genome assembly.
BOX2. WHOSE GENOME IS IT ANYWAY?
