THE PROBLEM OF NUCLEOTIDE SEQUENCE IN DEOXYRIBONUCLEIC ACIDS 83 



base composition, were vastly different with regard to the ahgnment of 

 their nucleotide constituents (Table I), the more important and interesting 

 uses of this method bear on other aspects of the sequence problem. These 

 relate, in fact, to the existence of sequential patterns that all deoxyribonu- 

 cleic acids examined heretofore appear to have in common. 



The first consequence of our studies may be the statement that in no 

 case can the nucleotide sequence be predicted from the knowledge of the 

 frequency of the individual constituents. The discarded tetranucleotide 

 hypothesis is a primitive example of an attempt to make such a prediction ; 

 I have dealt with this a long time ago [i]. This is, incidentally, quite similar 

 to what one would observe in a meaningful text : a certain letter may be 

 particularlv frequent; but this does not mean that it will occur, with 

 commensurate frequency, as a doublet. We shall probably find, in general, 

 that the symbols constituting a text are not arranged at random, if this 

 term is understood to refer to an alignment governed only by the relative 

 frequency of the elements, once the requirement of large numbers is 

 fulfilled. " 



This brings up the problem of the non-randomness of the nucleotide 

 sequence in deoxyribonucleic acids which we have discussed fully in a 

 previous paper [21]. Whether strict randomness is conceivable in a living 

 system capable of orderly replication, at any rate on the level of macro- 

 molecular events, is a question which it would be out of place to discuss 

 here. The entire complex of randomness in biology would certainly deserve 

 a more profound examination than I am equipped to perform. In any 

 event, the conclusion that in no case investigated the pattern of nucleotide 

 distribution was that to be expected from a random arrangement of 

 constituents [21, 35, 37, 40] may contribute to the assurance that in 

 studying the nucleotide sequence in a deoxyribonucleic acid we are, indeed, 

 dealing with some sort of meaningful message. If the array of constituents 

 in a biologically functional polynucleotide has been compared to a giant 

 throw of dice, we must conclude that it is a throw of loaded dice. 



Another consequence of our studies on deoxyribonucleic acids of 

 animal and plant origin [21, 37] is the conclusion that at least 70*^0 of the 

 pyrimidines occur as oligonucleotide tracts containing three or more 

 pyrimidines in a row; and a corresponding statement must, owing to the 

 equality relationship, apply also to the purines. 



POSSIBILITIES OF CLUSTER ANALYSIS 



It is quite clear that the various solitary nucleotides and the small 

 bunched oligonucleotides, such as doublets or triplets, will vary in fre- 

 quency, i.e. in relative quantity, according to the plan governing the nucleo- 

 tide sequence of the particular deoxyribonucleic acid; but they are most 



