The Release 5.1 annotation of Drosophila melanogaster heterochromatin

Abstract

The repetitive DNA that constitutes most of the heterochromatic regions of metazoan genomes has hindered the comprehensive analysis of gene content and other functions. We have generated a detailed computational and manual annotation of 24 megabases of heterochromatic sequence in the Release 5 Drosophila melanogaster genome sequence. The heterochromatin contains a minimum of 230 to 254 protein-coding genes, which are conserved in other Drosophilids and more diverged species, as well as 32 pseudogenes and 13 noncoding RNAs. Improved methods revealed that more than 77% of this heterochromatin sequence, including introns and intergenic regions, is composed of fragmented and nested transposable elements and other repeated DNAs. Drosophila heterochromatin contains "islands" of highly conserved genes embedded in these "oceans" of complex repeats, which may require special expression and splicing mechanisms.

Fig. 1

Computational pipeline results used for the Release 5.1 annotation. An Apollo (36) screenshot of the evidence (black region) and Release 5.1 annotations (light blue region) for scaffold CP000218, which was produced by merging, extending, and finishing the Release 3 WGS scaffolds (AABU01002750 and AABU010027230) (9). New cDNA evidence was used to merge the Release 3.2b annotations CG40067, CG17443, and CG40109 into one Release 5.1 gene (CG41520) and to identify an alternative exon for CG41250-RB (asterisk). ncRNA CG40375 is shown on the opposite strand to illustrate that it is nested within CG41520. CG40388 represents a Release 3.2b gene that is now annotated as a TE fragment. Complete annotation and evidence are shown in (11).

Fig. 2

Evidence of D. melanogaster heterochromatin protein-coding gene annotations. Venn diagrams show the percentage of protein-coding genes supported by gene prediction (pink), EST or cDNA (yellow), and/or BLASTX/TBLASTN comparative genomic evidence (blue). (A) Multi-exon genes are likely to be complete, whereas (B) single-exon genes are likely to represent genes that are fragmented across multiple scaffolds. The number of genes measured for each class are indicated, as well as the number and percent of genes with putative orthologs in melanogaster group species, nonmelanogaster group Drosophilids, or other insect species.

Fig. 3

Comparative analysis summary for D. melanogaster heterochromatin genes. (A) Number of heterochromatin protein-coding genes with a predicted ortholog in a given species, ordered (left to right) by increasing evolutionary distance from D. melanogaster.(B) Frequency histograms showing the percentage of heterochromatin protein-coding genes with a predicted ortholog in the 16 insect species tested. (C)Scatter plots of intron lengths(bp) for euchromatin and heterochromatin protein-coding gene introns conserved in either D. erecta or D. virilis. Each data point refers to a single conserved intron. Correlation coefficients (r) for intron lengths are indicated.

Fig. 4

Density of repeat and gene features across the heterochromatin. The average percentages of indicated annotation types are shown for each chromosome region (total length in boxes below the x axis). Euchromatin is an average of the noncentric regions of arms 2, 3, and X only. 2Lh, 2Rh, 3Lh, 3Rh, and Xh describe heterochromatic regions that are contiguous with the chromosome arms, whereas the Het regions are mapped to arms and ordered, but not necessarily in the correct orientation (8, 9). The average percentages of sequences for LTR- and LINE-like retrotransposons (red), DNA transposons (green), other and unknown repeats (blue), and TRF tandem repeats and satellite sequences (gray) are indicated. The average number of genes per 100 kb is shown above each histogram.

Fig. 5

Overrepresented GO terms in heterochromatin versus euchromatin genes. The percentages of GO molecular-function domains for genes in heterochomatin (red) and euchromatin (blue) are shown. Numbers in parentheses indicate the actual number of domains in heterochromatin. P-value significance scores are shown to the right. Complete GO analyses are presented in the SOM text.