Eukaryotic genomes were thought to be organized into highly stable chromosomes, with little genetic variation between cells. In the past, various studies have highlighted that besides the canonical chromosomes, eukaryotic genomes contain a plethora of non-chromosomal, circular DNA elements. While modern sequencing technologies provide the potential to investigate their impact, the research community is missing computational methods to deal with the vast amount of data generated by these techniques. As a result, the role of these circular DNA elements in the eukaryotic genomes remains to be established. Here, I present a new probabilistic method to detect DNA circles from shortread sequencing data, as well as 3 studies exploring the impact of circular DNA on eukaryotes.
The first chapter provides an overview about the circular DNA field. It serves as a brief introduction to the most intriguing topics behind the biology of DNA circles. The second chapter introduces Circle-Map, a new probabilistic method to detect circular DNA. The method guides the realignment of partially mapped reads using discordant reads in the vicinity and it is markedly more accurate than existing approaches. In the third chapter, we applied Circle- Map to study the segregation of endogenous circles as yeast cells divide and age. We found that young cells have highly diverse circular DNA populations, but this diversity is substantially diminished in their aged counterparts. Interestingly, DNA circles maintained in the yeast populations were characterized by replication origins. Furthermore, we observed that circles can have flexible inheritance when grown under selective pressure.
The fourth chapter presents the first assessment of chromosome derived DNA circles in the healthy human genome. We purified and sequenced circular DNA from 16 healthy individuals and found 100,000 unique circular DNA elements. We found that the majority circles are small, but some can be large enough (>25kb) to span full genes. Gene-rich chromosomes form DNA circles more frequently, and we show a fraction of DNA circles are transcriptionally active. The fifth chapter is a perspective outlining that if circular DNAs are transcriptionally active they could be a strong confounding factor in circular RNA sequencing studies, as transcription from circular DNAs have the potential to produce identical signals to those used to detect RNA circles.
In the last chapter we studied the condensed and less repeat-rich pigeon genome to address whether the size of the genome and its repetitive content affect the number of DNA circles. We detected more than 30,000 unique DNA circles by sequencing several pigeon breeds, and we discovered that non-flying pigeon breeds carry more circular DNA than flying pigeons. Furthermore, we compared the repetitive circular DNA levels between humans and pigeons, finding that DNA circles from repeats are in similar proportions to the repetitive content of the genome of origin.