Genomics

An active DNA transposon family in rice. Jiang, N. et al. Nature 421, 163–167 (2003)

The plant MITE mPing is mobilized in anther culture. Kikuchi, K. et al. Nature 421, 167–170 (2003)

Mobilization of a transposon in the rice genome. Nakazaki, T. et al. Nature 421, 170–172 (2003)

These three papers report the discovery of a family of active miniature inverted-repeat transposable elements (MITEs) in rice, which the authors call miniature Ping (mPing). This is an important finding as mPing is the first active MITE to be identified in any organism and the first active DNA transposon to be found in rice. A key feature of this transposon is that it reinserts with high frequency into low-copy coding regions of the rice genome. mPing mobilisation appears to be induced by stress such as gamma-radiation or cell-culture and relies on transposase activity provided in trans, probably to varying degrees by the related DNA transposons Ping and Pong. mPing might be suitable for use in developing gene-tagging programmes in rice. Such a programme could lead to the identification of genes controlling economically important rice traits and, therefore, facilitate the improvement of rice cultivars.

Human Genetics

Genetic structure of human populations. Rosenberg, N. A. et al. Science 298, 2381–2385 (2002)

The pattern of selection and migration of our ancestors is recorded in our DNA. In the largest surveys of its kind, the authors have genotyped 1,056 individuals from 52 populations by using 377 autosomal microsatellite markers. Using a statistical analysis that clusters individuals solely on the basis of their genetic similarity, the authors were able to assign the individuals sampled to the five main geographical regions and to subclusters that often corresponded to smaller populations. This work also confirms that most genetic variation exists within populations (93–95%) rather than between populations (3–5%).

Bioinformatics

Modeling the percolation of annotation errors in a database of protein sequences. Gilks, W. R. et al. Bioinformatics 18, 1641–1649 (2002)

Functional annotation of protein databases often relies on sequence homology, the functional annotation of which might have also been determined on the same basis. Gilks et al. refer to this possible chain of misannotations as 'error percolation' and develop a way to model the annotation quality, which clearly shows that this iterative approach quickly leads to decreased database quality. The authors use this as a starting point to build a scoring mechanism to qualitatively evaluate homology-based annotation.