201847a0Nature2014921196402228478470028-0836196410.1038/201847a0ukNatureNatureNATUREnatureNature is a weekly international journal publishing the finest peer-reviewed research in all fields of science and technology on the basis of its originality, importance, interdisciplinary interest, timeliness, accessibility, elegance and surprising conclusions. Nature also provides rapid, authoritative, insightful and arresting news and interpretation of topical and coming trends affecting science, scientists and the wider public./nature/journal/v201/n4921issueJournal homeArchiveCurrent issueAdvance online publicationPrivacy policySubscribeNature Publishing GroupCurrent issue201847a0A Preliminary Estimate of the Number of Human Genes
AU  - VOGEL, F.Institut fur Anthropologie und Humangenetik, University of Heidelberg, Germany.RECENT results of molecular genetics enable us to estimate the number of human genes, if certain assumptions are made. The following data are available: 
(1) The [alpha]-chain of human haemoglobin contains 141, the [beta]-chain contains 146 amino-acids, corresponding to a molecular weight of about 17,000 each1. Assuming a triplet code2,3 this means that the [alpha]- and [beta]-chains are determined by 423 and 438 nucleotide pairs, respectively. According to /`Svedberg's law/'4, many proteins consist of sub-units of the same order of magnitude (molecular weight of about 17,500). Hence, the assumption seems to be warranted that one average structural geno might have a length of about 450 nucleotide pairs.
(2) The weight of one haploid human chromosome set in human spermatozoa is about 2.72 [times] 10-12 g. Granulocytes contain about 6.23 [times] 10-12 g; lymphocytes contain about 5.84 [times] -12 g (ref. 5). Extensive examinations have shown that the DKA. content is constant in all resting cells of one species, which have the same number of chromosome sets, and depends on the degree of polyploidy5,6. The assumption seems to be justified that most of the DNA works as genetic material, even if in some cells minor fractions with other functions might possibly be present7. In the following calculations the total amount of DNA in a haploid human chromosome set is estimated to be about 3 [times] 10-12 g.
(3) Usually the genetic variants of human haemoglobins differ in one amino-acid substitution only1,8. One structural gene can only produce one single type of genetically determined polypeptide chain. As much as we know, this applies for other genetically determined proteins as well. This means that the genetic information for these structural genes can only be present once. Any degree of polyteny for these loci in the germ cells is highly unlikely. As has been mentioned, however, the DNA content of diploid cells is about twice the content of (haploid) spermatozoa. We assume that the total genetic information is only present once.
As can easily be calculated, a nucleotide pair with one adenine, one thymine, two deoxyribose, and two phosphate residues has a weight of 1.025  10?21 g, whereas a nucleotide pair with guanine and cytosine residues weighs 1.027  10?21 g. The difference can be neglected, and 1  10?21 g can be accepted as a good approximation. Hence, the total haploid chromosome set (3  10?12 g) contains  3  109 nucleotide pairs. Assuming 450 nucleotide pairs for one average structural gene, we arrive at an estimate of  6.7  106 structural genes per haploid chromosome set.
This estimate is disturbingly high indeed, and hence a different way of reasoning might be appropriate. In the giant chromosomes of certain Diptera which show a high degree of polyteny, evidence has been brought forward that the bands are functional units of protein synthesis. Within these bands one single chromatid thread has an average length of  50,000 nucleotide pairs9. This estimate of a ' gene -length' is about a hundred times higher than the estimate based on the protein data, and would lead to  6  104 units of this type for the haploid human genome. It is in much better accordance with the very crude assumptions made so far10. The question as to the difference with a factor of about 100 between the two estimates remains to be answered.
In my opinion, the answer might be that the systems of higher order which are connected with structural genes in operons and regulate their activity11 might occupy a much larger part of the genetic material than the structural genes which produce the polypeptide chains required for synthesis of enzymes and other functional proteins. The argument will be presented in greater detail later12.Hilschmann, , N., Blut, 8, 433 (1961).Nirenberg, , M. W., Sci. Amer., 208, 80 (1963).PubMedISIChemPortWittmann, , H. G., Naturwiss., 50, 76 (1963).ArticleISIChemPortKarlson, , P., Kurzes Lehrbuch der Biochemie (Georg-Thieme-Verlag, Stuttgart, 1962).Mirsky, , A. E., and Ris, , H., J. Gen. Physiol., 34, 451 (1951).ArticlePubMedISIChemPortVendrely, , R., in Nucleic Acids, edit. by Chargaff, E., and Davidson, J., 2, 155 (1955).Sampson, , M., Katoh, , A., Hotta, , Y., and Stern, , H., Proc. U.S. Nat. Acad. Sci., 50, 459 (1963).ChemPortBaglioni, , C., in Taylor, Molecular Genetics, Part 1, 405 (Academic Press, New York and London, 1963).Beermann, , W., Protoplasmatologia. 6, D, 1 (Springer Verlag, Wien, 1962).Neel, , J. V., and Schull, , W. J. (Univ. Chicago Press, 1954).Jacob, , F., and Monod, , J., in Biological Organization of the Cellular and Supercellular Level, edit. by Harris, R. J. C. (Academic Press, New York and London, 1963).Vogel, , F., Z. menschl. Vererbungs- u. Konstitutionslehre (in the press).