RS-centric intron expansion involves both size and position effects
To look into distinctive effects of TEs and SSs on intron size and position parameters, we divided introns into four basic classes: TS (both RSs), T (TEs), S (SSs), and N (neither TE nor SS). We focused on three essential intron features: fraction, length, and relative position in a gene. We made the following observations (Fig. 3). First, when plotting the percentage of introns in the four classes, we found that the pattern is rather heterogeneous, ie, the primates, the large mammals, and platypus are grouped together in a pattern of T > N > TS > S, showing a transposon-dominant pattern, so is opossum that has a pattern of T > TS > N > S. Second, mouse and rat form their own group, as it is noticed that both have more satellite sequences than other mammals: TS > N > T > S. Third, aside from the dominant TS-free group or N, guinea pig (N > T > TS > S), frog (N > T > TS > S), and chicken (N > T > TS > S) all have more transposons in their introns than satellites. Fourth, anole and zebrafish have a pattern of N > TS > T > S, in a similar path as compared to mouse and rat regardless of N. If we pick a single most abundant RS-containing intron group, TS, T, S, and N, for a species, the fractions are 39.6%, 52.7%, 12.8%, and 72% in mouse, platypus, anole, and chicken, respectively.
We also investigated the size relevance of introns according to two simple size intervals: ≤1000 bp and >1000 bp. Obviously, the absolute majority of introns in N are small, ≤1000 bp, as opposed to the fact that the greater majority of introns in TS and T are larger, >1000 bp. When examining the median length, we found that intron length increase is correlated with the complexity of RS insertions: TS > T > S > N (Fig. 4). We also observed that the TS intron group tends to be near the 5′-end of genes as opposed to the N intron group that tends to be near the 3′-end of the genes in primates, large mammals, rodents, opossum, and frog, as well as that the TS intron group tends to be near the 5′-end of the genes in platypus, chicken, and anole (Fig. 5). The extremely biased distributions are seen in mouse, where the transposon-rich introns tend to be near the 3′-end, and in zebrafish, where all four intron groups show no significant bias.
We further examined both length and position effects for four selected transposons: LTR, LINE, SINE, and DNA. Their intron length medians rank as LTR > DNA > LINE > SINE in the primates, the large mammals, and opossum (Fig. 6). In the three rodents, mouse and rat form a unique league themselves with a length order of DNA > LINE > LTR > SINE, but guinea pig stands alone with a similar pattern to other non-rodent mammals: LTR > DNA > LINE > SINE. In addition, the platypus introns with LTR or DNA transposons tend to be larger in size, in comparison with those of LINE- or SINE-containing introns. In contrast, the chicken introns with LINE tend to be smaller, when compared to those with SINE, DNA or LTR. There are other independent patterns such as LTR > SINE > LINE > DNA and LTR > LINE > SINE > DNA in frog and zebrafish, respectively. An exception is unique to anole, where the order becomes LINE > SINE > DNA when LTR is absent. The most likely reason is the lack of well-classified LTR consensus in the RepeatMasker default library due to high diversity of transposable elements in anole, especially when compared to mammals.17 In the primates, the large mammals, and guinea pig, the median position index ranks as LTR < DNA < LINE < 0, and the introns with SINEs in cow, panda, horse, human, and guinea pig have a slight bias toward 5′-end (data not shown). In both mouse and rat, the introns with DNA transposons have the most 5′-end biases and those with SINEs have the least 5′-end biases. In the two primitive mammals, opossum and platypus, their LTRs and DNA transposons tend to be inserted into introns near the 5′-end. The chicken introns harbouring LTRs or DNA transposons have a stronger bias toward insertions at the 5′-end than those with LINE. The order of the median intron position index for anole is LINE < SINE < DNA < 0. The positional preference for the frog introns is the proximity of 5′-end but that of DNA transposon- containing introns is the weakest. In zebrafish, introns with LINE, SINE or LTR have a stronger 5′-end preference, and those with LTR have the least bias.