08 ERWIN CHARGAFF 



learn something about the structural principles that came into play. I 

 shall first discuss some of the basic concepts that must be considered. 



Remarks on the conceptual basis of sequence analysis 



THE PROBLEM OF HOMOGENEITY 



What determines the composition of a given biological polymer has 

 been much discussed in the recent past, at any rate in regard to proteins 

 and nucleic acids. Almost no attention has been paid, as has been pointed 

 out recently [2], to the presumably equally rigid control of the composition 

 of other cell-specific polymers, such as the polysaccharides or even certain 

 macromolecular lipids. The usual expedient, in which I wish I could 

 concur with greater enthusiasm [3], has been the formulation of some 

 sort of template which obviously makes up in versatility for what it lacks 

 in concreteness. This Disneyland of templates and pools and feedbacks, 

 presided over by never-smiling augurs, will make us the laughing stock of 

 later times. 



What determines the size of a given biological polymer has, on the 

 other hand, been neither determined nor even much discussed. The cell 

 will in many instances contain, or at least it will be made to yield, com- 

 pounds belonging to the same class, but very much dilTerent in size. Pro- 

 teins, the best investigated group of cellular macromolecules, vary in size 

 over a considerable range, from something like 12 000 to many millions [4]. 

 Deoxyribonucleic acid preparations appear, in contrast, to fall into two 

 principal groups of molecular weight: (a) of 6 to 8 million; (b) of 12 to 16 

 million [5].* Such distinctions are probably not of great heuristic value, 

 since the cell, which is not simply a bag containing many chemicals of 

 different size and quality, must impress on its components a pattern of 

 multiple associations which it is not possible even to define at present. 



Dissimilarity in size is, however, less of a problem for the student of 

 the chemistry of deoxyribonucleic acids than is the possible heterogeneity 

 of his specimens as regards their composition and sequence characteristics. 

 It is, in fact, possible to divide a preparation of total deoxyribonucleic 

 acid into a series of fractions of graded composition (starting with a, some- 

 times extreme, GC-type [7] and going to a very marked AT-type) which, 

 however, all still exhibit the pairing regularities [8] ; but the real meaning 

 of such a fractionation is far from clear. It could be, as was pointed out by 

 us [9], that these fractionated preparations actually represent some form 

 of sub-units of a large aggregate, though it is not easy to describe the nature 



* An interesting exception — a polydeoxyribonucleotide of mol. wt. 10 000 to 

 12 000 — has been described recently as a component of crystalline cytochrome 

 62 [6]. It is noteworthy that this small polymer of only about 40 nucleotides still 

 exhibits the pairing principles [i] mentioned above. 



