The Human Genome Project, Genomics, and Proteomics

The Human Genome Project (1990 to 2003) was an international effort to determine the sequence of the roughly 3.2 billion base pairs of the human genome; genomics studies whole genomes, proteomics studies the full set of proteins (the proteome), and genetic testing uses this knowledge to detect inherited variants. In brief: the human genome contains about 20 000 protein-coding genes, but these make up only about 1 to 2 percent of the DNA. The number of different proteins is much larger than the number of genes, because of alternative splicing and chemical modification. Any two people share about 99.9 percent of their DNA sequence.

The genome project opened the way to personalized medicine and changed biological research. This guide describes how the project was carried out and what it found, explains genomics and proteomics, the main types of genetic variation, genetic testing and its uses, works through a calculation about genome size and data, and corrects common mistakes.

What was the Human Genome Project?

The Human Genome Project (HGP) was a public effort, led by the US National Institutes of Health and the Department of Energy and carried out with partners in the UK, Japan, France, Germany and China. It began officially in 1990, with the aim of determining the order of all base pairs and identifying all genes, at an expected cost of 3 billion dollars. A private company, Celera Genomics, led by Craig Venter, began a competing effort in 1998 with a faster approach (whole-genome shotgun sequencing). The two groups announced a working draft in June 2000 and published their drafts in February 2001. The project declared the sequence complete, covering 92 percent of the genome, in April 2003, 50 years after the discovery of the structure of DNA. In 2022, the Telomere-to-Telomere (T2T) consortium published the first truly complete human genome sequence, filling the remaining gaps.

The project also set principles for sharing data freely and funded studies of the ethical, legal and social implications of genome research (ELSI), which received about 3 to 5 percent of the budget.

What did the human genome reveal?

• Size: about 3.2 billion base pairs in a haploid set, spread over 23 chromosomes. The smallest chromosome (21) is about 47 million base pairs, and the largest (1) is about 248 million.
• Genes: about 20 000 protein-coding genes, fewer than the 100 000 or more predicted before the project. A fruit fly has about 14 000 genes, a nematode worm about 20 000, and rice about 40 000. The number of genes does not correlate with complexity.
• Non-coding DNA: only about 1 to 2 percent codes for proteins. A large share is made of repetitive sequences, including transposable elements (about 45 percent of the genome). Other regions are genes for functional RNAs and regulatory regions such as promoters and enhancers. The ENCODE project reported in 2012 that about 80 percent of the genome has some biochemical activity, but how much of this is functionally important is debated.
• Variation: two people differ at about 1 in every 1 000 bases, a total of roughly 3 to 4 million differences. Most are single nucleotide polymorphisms (SNPs). There are also insertions, deletions and copy number variants.
• Similarity to other species: humans share about 98 to 99 percent of sequence with chimpanzees in aligned regions, and genes with the same function are found in distant species (about 60 percent of fruit fly genes have a human counterpart).

What is genomics?

Genomics is the study of the structure, function, evolution and mapping of whole genomes. It includes:

• Structural genomics: determining the sequence and organization of genomes (as in the HGP).
• Functional genomics: finding what genes and regulatory elements do, by studying gene expression (the transcriptome, using RNA sequencing), gene knockouts and CRISPR screens.
• Comparative genomics: comparing genomes between species to find conserved regions (likely important) and to reconstruct evolution.
• Population genomics: studying variation in groups of people. Projects such as the 1000 Genomes Project and UK Biobank, with half a million participants, relate genome variants to traits and diseases through genome-wide association studies (GWAS).
• Metagenomics: sequencing all DNA in a sample from the environment or the body (the microbiome).

Bioinformatics, the use of computers to store, search and analyze biological data, is essential, as the data are so large.

What is proteomics?

The proteome is the entire set of proteins made by an organism, tissue or cell at a given time. Proteomics studies their identity, amount, modifications, locations and interactions. It is more complex than genomics, since:

• The proteome varies between cell types and over time, while the genome is nearly the same in all cells.
• One gene can give several proteins by alternative splicing, and proteins are modified after translation (phosphorylation, glycosylation and so on). The human proteome may contain over a million protein forms from about 20 000 genes.
• The function of a protein depends on its three-dimensional shape and its interactions with other molecules.

The main tools are mass spectrometry, which identifies proteins by the masses of their fragments, protein arrays and, for structures, X-ray crystallography, NMR and cryo-electron microscopy. Machine-learning systems such as AlphaFold, introduced by DeepMind in 2020, predict the three-dimensional structures of proteins from their amino acid sequence with high accuracy, and a database with predicted structures for over 200 million proteins was released in 2022. Proteomics is used to find biomarkers for disease (for example, proteins in blood that indicate cancer or heart damage) and to identify drug targets.

Genomics and proteomics compared

What is genetic testing?

Genetic tests analyze DNA, RNA, chromosomes or proteins to find changes linked to disease. The main types are:

• Diagnostic testing: confirms or rules out a suspected genetic condition (such as cystic fibrosis or Huntington’s disease).
• Carrier testing: finds people who carry one copy of a recessive disease allele (for example, for cystic fibrosis, sickle-cell disease and Tay-Sachs disease), so couples can learn their chance of having an affected child.
• Prenatal and newborn screening: for chromosome conditions and metabolic diseases such as phenylketonuria (PKU), where early treatment with a special diet prevents harm.
• Predictive and risk testing: estimates the chance of developing a disease later, e.g. BRCA1/BRCA2 variants (increased risk of breast and ovarian cancer). A polygenic risk score combines the effects of many variants.
• Pharmacogenomic testing: predicts the response to drugs, so the choice and dose of a drug can be adjusted.
• Direct-to-consumer tests: sold for ancestry and health; their reliability and interpretation vary.

Most common diseases such as heart disease and diabetes are influenced by many genes together with lifestyle, so a test gives a probability, not a certainty. A “positive” result for a single-gene disease such as Huntington’s disease predicts the disease with near certainty, but a BRCA1 variant means increased risk, not that the cancer will occur.

Worked example: genome size, sequencing coverage and data

A laboratory sequences a human genome with short reads of 150 bases, aiming for 30× coverage (each base is read 30 times on average, to detect errors and to distinguish true variants).

1. Total bases to read: 3.2 × 10⁹ × 30 = 9.6 × 10¹⁰ bases (96 billion).
2. Number of reads: 9.6 × 10¹⁰ ÷ 150 = 6.4 × 10⁸ reads (640 million).
3. Data storage: if each base takes about 2 bytes in the raw file with quality scores (1 byte for the base and 1 for its quality), the raw data are 9.6 × 10¹⁰ × 2 ≈ 1.9 × 10¹¹ bytes, about 190 gigabytes (before compression). The final list of variants is far smaller.
4. Expected variants: at 1 difference per 1 000 bases relative to the reference genome: 3.2 × 10⁹ ÷ 1 000 = 3.2 million variants. A typical genome includes 4 to 5 million.
5. Information content: with 2 bits per base, 3.2 × 10⁹ × 2 = 6.4 × 10⁹ bits = 0.8 gigabytes for one haploid genome. This is similar to a CD-ROM, and shows that the information contained in the genome is small compared with the data of the experiment.

The numbers explain why genome sequencing needs powerful computers and why the costs of data storage and analysis are a large part of the total expense. The figures are rounded.

What are the ethical and social issues?

• Privacy and confidentiality: genetic information is personal and also reveals information about relatives. Laws such as the US Genetic Information Nondiscrimination Act (2008) forbid discrimination in health insurance and employment, but have gaps (for example, in life insurance).
• Consent and the right not to know: people may not wish to learn that they carry a gene for an incurable disease.
• Incidental findings: a test may reveal something other than what was looked for.
• Equity: most genomic studies have used participants of European ancestry, so tools may work less well for other groups, and access to genomic medicine is unequal.
• Interpretation: many variants have unknown significance, and risk scores can be misunderstood. Race is not a precise genetic category, as human genetic variation is mostly within groups.
• Ownership and data sharing: who controls genomic data and benefits from it?

How did the project develop?

Frederick Sanger’s sequencing method (1977) and the first sequenced genome, of the bacteriophage phiX174 (1977), laid the foundations. Robert Sinsheimer, Renato Dulbecco and Charles DeLisi helped to launch the HGP in the 1980s. Francis Collins led the public project from 1993, and Craig Venter led Celera. President Bill Clinton and Prime Minister Tony Blair announced the draft in June 2000. The cost of sequencing a human genome fell from about 100 million dollars in 2001 to under 1 000 dollars in the 2020s, a decrease much faster than the decline in the cost of computing.

Which mistakes are common?

Mistake 1: Thinking that the genome contains only genes

Genes that code for proteins are only 1 to 2 percent of the DNA. The rest includes regulatory regions, RNA genes and repeats.

Mistake 2: Saying that one gene makes one protein

Alternative splicing and modifications give many protein forms from one gene.

Mistake 3: Believing that the genome sequence is the “blueprint” that determines everything

Gene activity, the environment and chance also affect traits.

Mistake 4: Treating a genetic test result as a certain prediction

Except for a few single-gene diseases, the result shows a risk, not a certainty.