TEMP: a computational method for analyzing transposable element polymorphism in populations.
Bottom Line: TEMP also performs well on whole-genome human data derived from the 1000 Genomes Project.We applied TEMP to characterize the TE frequencies in a wild Drosophila melanogaster population and study the inheritance patterns of TEs during hybrid dysgenesis.We also identified sequence signatures of TE insertion and possible molecular effects of TE movements, such as altered gene expression and piRNA production.
Affiliation: Program in Bioinformatics and Integrative Biology, Department of Biochemistry and Molecular Pharmacology.Show MeSH
Related in: MedlinePlus
Mentions: We computed the frequency change for each parental TE insertion (defined as the insertions whose frequencies exceed 10% in at least one of the parental populations) (Figure 3a). As expected, the vast majority of parental insertions have negative but close to zero frequency changes in the F1 population (Figure 3a), suggesting that they were under weak purifying selection. The most critical challenge facing the hybrid dysgenic flies was coping with hyperactive transpositions and any trait that helped suppress TE mobilization could be potentially selected for. Insertion of a TE into piRNA clusters can lead to production of piRNAs whose sequences are complementary to the TE, and these piRNAs can in turn silence the corresponding transposon genome-wide (29–31). Indeed, among insertions whose frequencies increased by 30% or more in the F1 population (FC ≥ 0.3), there were more insertions residing within piRNA clusters than expected (P-value = 5.38E-4, hypergeometric test). In contrast, among insertions with FC ≤ −0.3, there were fewer of them than expected in piRNA clusters (P-value = 8.02E-5, hypergeometric test) (Figure 3a). We also analyzed the germline DNA isolated from the ovaries of the second-generation progenies (F2) produced by backcrossing F1 dysgenic females to w1 males. Again, we computed the FC for each parental TE insertion, i.e. those insertions whose frequency exceeded 10% in either F1 or w1. These F2 females did not suffer from hyperactive TE movement; accordingly, our data support that there were fewer insertions under negative selection than their parents (14.05% with FC ≤ −0.3 in F2 versus 19.09% in F1; P-value = 4.90E-5, X2-test). Moreover, there was no enrichment for TE insertions in piRNA clusters (P-value = 0.53, hypergeometric test), consistent with the notion that such insertions would not confer significant selective advantages in non-dysgenic individuals (Figure 3b). We note that according to the Wright–Fisher model with a population size of 100 (200 chromosomes), the probability of FC ≤ −0.3 or ≥ 0.3 or more extreme is smaller than 1E-15 (Supplementary Table S6). Therefore, the sites with FC ≤ −0.3 or ≥ 0.3 are likely under selective pressure. Moreover, at 20X sequencing depth TEMP's FDR is 1.17% for sites with frequency at 0.3 (Figure 2e). Therefore most of the sites with FC ≤ −0.3 or ≥ 0.3 represent actual change, not detection error.
Affiliation: Program in Bioinformatics and Integrative Biology, Department of Biochemistry and Molecular Pharmacology.