Under-representation of repetitive sequences in whole-genome shotgun sequence databases: an illustration using a recently acquired transposable element.

Abstract

It is widely accepted in a conceptual framework that repetitive sequences, especially those with high sequence homogeneity among copies, tend to be under-represented in whole-genome shotgun sequence databases, because of the difficulty of assembling sequence reads into contigs. Although this is easily inferred, there is no quantitative illustration of this phenomenon. An example using a currently used database is expected to contribute to the intuitive understanding of how serious the under-representation is. The present study provides the first quantitative example (in the case of 16 copies of virtually identical, 4.7-kb sequences in a genome of 7 × 10 (8) bp) by comparing the results of BLAST searches of a sequence database (contig N50; 9.8 kb) with those of Southern blot analysis of genomic DNA. This has revealed that the internal regions of the repetitive sequences are under-represented to a striking extent.