Happy birthday human genome sequence!

The draft sequences of the human genome were published fifteen years ago. The International Human Genome Project (IGHP) published its draft sequence in Nature on Feb. 15, 2001 (Lander et al., 2001) and Celera Genomics published its draft sequence in Science on Feb. 16, 2001 (Venter et al., 2001).¹

For me the timing was perfect since I was scheduled to give a Journal Club talk on March 16th and you could hardly ask for a better topic.

How much DNA was sequenced?

The IHGP sequenced about 2.6 Gb of DNA then added additional sequences from public databases to get to a total of 2.84 Gb of contiguous nucleotide sequence. This corresponds to almost 90% of the genome.

The Celera sequence was quite a bit less but it was supplemented with the IHGP sequences that were available in public databases prior to publication of the draft sequence. The total amount of genome sequence in the Celera paper was reported as 2.91 Gb or about 91% of the genome. Some of these reported sequences are just strings of N's so the actual amount of the genome sequence in both cases is closer to 89% (IHGP) and 83% (Celera) (Aach et al., 2001).

How much junk?

When the idea of spending vast amounts of money to sequence the human genome was first raised in the late 1980s, there were several critics who pointed out that most of our genome is junk and sequencing junk DNA would be a waste of money. In spite of this criticism, it was decided that the project should go ahead. It would be good to see what this junk looked like and there might even be some useful knowledge in the junk DNA.

The International Human Genome Project reported that 50% of the genome consisted of repeat sequences, as expected. Most of them are derived from transposons. They devote 10.5 pages of their paper to discussing the nature of these repeats sequences and their evolution.

Repeats are often described as 'junk' and dismissed as uninteresting. However, they actually represent an extraordinary treasure trove of information about biological processes. The repeats consitute a rich paleontological record, holding crucial clues about evolutionary events and forces.

Lander et al. propose that a small number of these sequences, mostly Alu's could be adaptive.

Celera Genomics says that 35% of the genome consists of repeats. They get one paragraph (p. 1323).

How many genes did they find?

The International Human Genome Project uses a definition of "gene" that corresponds to the one I prefer; namely, "a DNA sequence that's transcribed." They take the time to point out that there are many genes that do not code for protein.

Although biologists often speak of a tight coupling between ‘genes and their encoded protein products’, it is important to remember that thousands of human genes produce noncoding RNAs (ncRNAs) as their ultimate product. There are several major classes of ncRNA. (1) Transfer RNAs (tRNAs) are the adapters that translate the triplet nucleic acid code of RNA into the amino-acid sequence of proteins; (2) ribosomal RNAs (rRNAs) are also central to the translational machinery, and recent X-ray crystallography results strongly indicate that peptide bond formation is catalysed by rRNA, not protein; (3) small nucleolar RNAs (snoRNAs) are required for rRNA processing and base modification in the nucleolus; and (4) small nuclear RNAs (snRNAs) are critical components of spliceosomes, the large ribonucleoprotein (RNP) complexes that splice introns out of pre-mRNAs in the nucleus. Humans have both a major, U2 snRNA-dependent spliceosome that splices most introns, and a minor, U12 snRNA-dependent spliceosome that splices a rare class of introns that often have AT/AC dinucleotides at the splice sites instead of the canonical GT/AG splice site consensus.

Other ncRNAs include both RNAs of known biochemical function (such as telomerase RNA and the 7SL signal recognition particle RNA) and ncRNAs of enigmatic function (such as the large Xist transcript implicated in X dosage compensation250, or the small vault RNAs found in the bizarre vault ribonucleoprotein complex, which is three times the mass of the ribosome but has unknown function).

They then go on, over the next three pages, to discuss the number of genes in each of these categories and how closely they correspond to previous estimates.

In contrast, the Celera Genomics group defines a gene as "a locus of cotranscribed exons" and they don't even talk about genes for functional RNAs.

The IHGP group gives us a summary of previous estimates of the number of protein-coding genes in the human genome. Note the conclusion ... they were expecting about 30,000 genes based on reading the literature.

Previous estimates of human gene number. Although direct enumeration of human genes is only now becoming possible with the advent of the draft genome sequence, there have been many attempts in the past quarter of a century to estimate the number of genes indirectly. Early estimates based on reassociation kinetics estimated the mRNA complexity of typical vertebrate tissues to be 10,000–20,000, and were extrapolated to suggest around 40,000 for the entire genome. In the mid-1980s, Gilbert suggested that there might be about 100,000 genes, based on the approximate ratio of the size of a typical gene (~3 × 10⁴ bp) to the size of the genome (3 × 10⁹ bp). Although this was intended only as a back-of-the-envelope estimate, the pleasing roundness of the figure seems to have led to it being widely quoted and adopted in many textbooks. (W. Gilbert, personal communication). An estimate of 70,000–80,000 genes was made by extrapolating from the number of CpG islands and the frequency of their association with known genes.

As human sequence information has accumulated, it has been possible to derive estimates on the basis of sampling techniques. Such studies have sought to extrapolate from various types of data, including ESTs, mRNAs from known genes, cross-species genome comparisons and analysis of finished chromosomes. Estimates based on ESTs have varied widely, from 35,000 to 120,000 genes. Some of the discrepancy lies in differing estimates of the amount of contaminating genomic sequence in the EST collection and the extent to which multiple distinct ESTs correspond to a single gene. The most rigorous analyses exclude as spurious any ESTs that appear only once in the data set and carefully calibrate sensitivity and specificity. Such calculations consistently produce low estimates, in the region of 35,000.

Comparison of whole-genome shotgun sequence from the pufferfish T. nigroviridis with the human genome can be used to estimate the density of exons (detected as conserved sequences between fish and human). These analyses also suggest around 30,000 human genes.

Extrapolations have also been made from the gene counts for chromosomes 21 and 22, adjusted for differences in gene densities on these chromosomes, as inferred from EST mapping. These estimates are between 30,500 and 35,500, depending on the precise assumptions used.

The IHGP estimated that there were about 31,000 protein-coding genes and "this is consistent with most recent estimates based on sampling, which suggest a gene number of 30,000 - 35,000." Thus, the Lander et al. group were not surprised at the predicted number of genes because it fit right in with previous estimates. (They don't mention the estimate based on genetic load.)

If this is what the paper actually said, then why was there so much hype about the "low" number of genes? It's partly due to the Venter et al. paper from Celera Genomics. They were also aware of previous work by knowledgeable experts because they said (page 1317),

The number of protein-coding genes in mammals has been controversial from the outset. Initial estimates based on reassociation data placed it between 30,000 to 40,000, whereas later estimates from the brain were >100,000. More recent data from both the corporate and public sectors, based on extrapolations from EST, CpG island, and transcript density–based extrapolations, have not reduced this variance. The highest recent number of 142,634 genes emanates from a report from Incyte Pharmaceuticals, and is based on a combination of EST data and the association of ESTs with CpG islands. In stark contrast are three quite different, and much lower estimates: one of ∼35,000 genes derived with genome-wide EST data and sampling procedures in conjunction with chromosome 22 data; another of 28,000 to 34,000 genes derived with a comparative methodology involving sequence conservation between humans and the puffer fish Tetraodon nigroviridis; and a figure of 35,000 genes, which was derived simply by extrapolating from the density of 770 known and predicted genes in the 67 Mbp of chromosomes 21 and 22, to the approximately 3-Gbp euchromatic genome.

A few pages later (p. 1346) they discuss the genetic load estimate of Muller and Crow & Kimura, pointing out that these 30 year-old estimates predicted about 30,000 genes.

It's clear that there was a strong consensus in the literature that the human genome would have only about 30,000 genes. Celera announced that there were about 26,500 genes that were well-supported by multiple lines of evidence and about 12,000 more that had weaker support. Then they said,

8.2 The low gene number in humans

We have sequenced and assembled ∼95% of the euchromatic sequence of H. sapiens and used a new automated gene prediction method to produce a preliminary catalog of the human genes. This has provided a major surprise: We have found far fewer genes (26,000 to 38,000) than the earlier molecular predictions (50,000 to over 140,000). Whatever the reasons for this current disparity, only detailed annotation, comparative genomics (particularly using the Mus musculus genome), and careful molecular dissection of complex phenotypes will clarify this critical issue of the basic “parts list” of our genome.

This is what got all the attention in subsequent press releases and the myth of "surprise" continues to be repeated [see Nessa Carey doesn't understand junk DNA].

---

The announcement of the completed project was six months earlier on June 26, 2000. Here's the opening paragraph of the book I'm working on ...

“It is humbling for me and awe-inspiring to realize that we have caught the first glimpse of our own instruction book, previously known only to God.” Those were the words of Francis Collins, director of the International Human Genome Project, as he announced the draft sequence of the human genome on June 26, 2000. The other two men on the stage were Bill Clinton, President of the United States, and Craig Venter, president of Celera Genomics. One of those other men, President Clinton, was quite comfortable with the god language. The other one probably thought they were referring to him.

Aach, J., Bulyk, M. L., Church, G. M., Comander, J., Derti, A., and Shendure, J. (2001) Computational comparison of two draft sequences of the human genome. Nature, 409:856-859. [doi: 10.1038/35057055]

Lander, E., Linton, L., Birren, B., Nusbaum, C., Zody, M., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., Funke, R., Gage, D., Harris, K., Heaford, A., Howland, J., Kann, L., Lehoczky, J., LeVine, R., McEwan, P., McKernan, K., Meldrim, J., Mesirov, J., Miranda, C., Morris, W., Naylor, J., Raymond, C., Rosetti, M., Santos, R., Sheridan, A., Sougnez, C. et al. (2001) Initial sequencing and analysis of the human genome. Nature, 409:860-921. [doi: 10.1038/35057062]

Venter, J., Adams, M., Myers, E., Li, P., Mural, R., Sutton, G., Smith, H., Yandell, M., Evans, C., Holt, R., Gocayne, J., Amanatides, P., Ballew, R., Huson, D., Wortman, J., Zhang, Q., Kodira, C., Zheng, X., Chen, L., Skupski, M., Subramanian, G., Thomas, P., Zhang, J., Gabor Miklos, G., Nelson, C., Broder, S., Clark, A., Nadeau, J., McKusick, V., Zinder, N. et al. (2001) The sequence of the human genome. Science, 291:1304-1351. [doi: 10.1126/science.1058040]