Studies On The Length Distribution Of Protein Coding Sequence In Genomes And Genome Evolution

The paper is composed of three chapters:In the first chapter, by comparing yeast genome date of three different period, the variations of ORFs and protein coding sequences for recent years are analysed. The variations of short ORFs are instable. It is possible that the instability comes from the randomness of the DNA sequences. Simultaneously, some obvious characteristics of ORFs named with TY*(TY* ORFs) and ORFs named in other ways (* ORFs) are obtained and the variations of all the * ORFs in yeast genomes of 2001 and 2002 years are discussed.In the second chapter, the distribution of protein coding sequences' number with its length is analyzed in 30 genomes, which are from Eukarya, Bacteria and Archaea. According to the characteristics of these distributions, a hypothesis that this kind of distribution is ( a , B) distribution is proposed. Compared with other distribution models, the distribution of protein coding sequences' number with its length accordswith r (a , B) distribution. By studying the parameters a and B values of (a ,B) distribution, a distinct relation between the values of a and B and the genome evolution is obtained. The significance represented by the values of a and B is discussed and the result that Eukarya have a bias towards the longer genes is given. In the third chapter, the distribution of gene coding sequences' number with its length in 24 chromosomes of the human genome is analyzed. It is shown that these distributions have obvious similarity. By use of the r ( a , B) distribution fitting to the real distributions, the values of its parameter a are all smaller than 1. But the a values in other organisms (15 bacteria, 10 archaea, and 5 eukaryotes) are all larger than 1. Compared with these results, we argue that the length distribution of protein coding sequences in the human genome is also the r ( a ,B) distribution with a >1. By complementing some available date to the short sequences and fitting to the new date by the r ( a , B) distribution, a good fitting result is obtained and the values of the parameter a are between 1.19 and 1.85.