Transposable elements (TEs) are major constituents of plant genomes and contribute to their dynamics and evolution. The extent to which they contribute to local adaptation nevertheless remains to be fully elucidated. To better characterize the potential role of TEs in maize local adaptation, we aim at characterizing TE polymorphisms in a set of 7 maize inbred lines of contrasted origins. We have de novo assembled the genome sequences of these maize lines, and are now focusing on TE annotation and polymorphism discovery. Several softwares and pipelines are available to de novo annotate TEs in genomes. They rely on (i) approaches that use characteristic structural features of TE super-families to detect complete copies and/or (ii) approaches that use the repetitiveness of TE copies across the genome. These approaches are often completed by the detection of coding domains to clas- sify of TEs into families. Existing softwares have proven useful for TE annotation of small genomes of a few hundreds of megabases, but remain challenging to apply on larger genomes, mainly due to the amount of genomic data to analyze and to the structural complexity of these genomes. With a ˜2.1 Gb genome, 85% of which is occupied by TE sequences that are often nested within one another, maize is a typical average diploid angiosperm genome. Improving tools for maize TE genome annotation will therefore leverage TE annotation for a large amount of plant species. After reviewing existing approaches to de novo annotate TEs, we will discuss their advantages and limits to annotate TEs in the maize genome. We will then suggest ways to improve de novo TE annotations in such a genome