Abstract
In this paper we present the ADOPS (Automatic Detection Of Positively Selected Sites) software that is ideal for research projects involving the analysis of tens of genes. ADOPS is a novel software pipeline that is being implemented with the goal of providing an automatic and flexible tool for detecting positively selected sites given a set of unaligned nucleotide sequence data.

Abstract
Chromosomal inversions can originate from breakage and repair by non-homologous end-joining. Nevertheless, they can also originate from ectopic recombination between transposable elements located on the same chromosome inserted in opposite orientations. Here, we show that a MITE element (DAIBAM), previously involved in the origin of one Drosophila americana polymorphic inversion, is also involved in the origin of one fixed inversion between D. virilis and D. americana and another D. americana polymorphic inversion. Therefore, DAIBAM is responsible for at least 20% of the chromosomal rearrangements that are observed within and between species of the virilis phylad (D. virilis, D. lummei, D. novamexicana and D. americana), having thus played a significant role in the chromosomal evolution of this group of closely related species.

Abstract
Background: Comparative studies using hundreds of sequences can give a detailed picture of the evolution of a given gene family. Nevertheless, retrieving only the sequences of interest from public databases can be difficult, in particular, when working with highly divergent sequences. The difficulty increases substantially when one wants to include in the study sequences from many (or less well studied) species whose genomes are non-annotated or incompletely annotated. Methodology/Principal Findings: In this work we evaluate the usefulness of different approaches of gene retrieval and classification, using the distal-less (DLX) gene family as a test case. Furthermore, we evaluate whether the use of a large number of gene sequences from a wide range of animal species, the use of multiple alternative alignments, and the use of amino acids aligned with high confidence only, is enough to recover the accepted DLX evolutionary history. Conclusions/Significance: The canonical DLX homeobox gene sequence here derived, together with the characteristic amino acid variants here identified in the DLX homeodomain region, can be used to retrieve and classify DLX genes in a simple and efficient way. A program is made available that allows the easy retrieval of synteny information that can be used to classify gene sequences. Maximum likelihood trees using hundreds of sequences can be used for gene identification. Nevertheless, for the DLX case, the proposed DLX evolutionary is not recovered even when multiple alignment algorithms are used.

Abstract
Background: Although homeobox genes have been the subject of many studies, little is known about the main amino acid changes that occurred early in the evolution of genes belonging to different classes. Results: In this study, we report a method for the fast and efficient retrieval of sequences belonging to the ANTP (HOXL and NKL) and PRD classes. Furthermore, we look for diagnostic amino acid residues that can be used to distinguish HOXL, NKL and PRD genes. Conclusion: The reported protein features will facilitate the robust classification of homeobox genes from newly sequenced bilaterian genomes. Nevertheless, in non-bilaterian genomes our findings must be cautiously applied. In principle, as long as a good manually curated data set is available the approach here described can be applied to non-bilaterian organisms as well. Our results help focus experimental studies onto investigating the biochemical functions of key homeodomain residues in different gene classes.

Abstract
Background: Within Rosaceae, the RNase based gametophytic self-incompatibility (GSI) system has been studied at the molecular level in Maloideae and Prunus species that have been diverging for, at least, 32 million years. In order to understand RNase based GSI evolution within this family, comparative studies must be performed, using similar methodologies. Result: It is here shown that many features are shared between the two species groups such as levels of recombination at the S-RNase ( the S-pistil component) gene, and the rate at which new specificities arise. Nevertheless, important differences are found regarding the number of ancestral lineages and the degree of specificity sharing between closely related species. In Maloideae, about 17% of the amino acid positions at the S-RNase protein are found to be positively selected, and they occupy about 30% of the exposed protein surface. Positively selected amino acid sites are shown to be located on either side of the active site cleft, an observation that is compatible with current models of specificity determination. At positively selected amino acid sites, non-conservative changes are almost as frequent as conservative changes. There is no evidence that at these sites the most drastic amino acid changes may be more strongly selected. Conclusions: Many similarities are found between the GSI system of Prunus and Maloideae that are compatible with the single origin hypothesis for RNase based GSI. The presence of common features such as the location of positively selected amino acid sites and lysine residues that may be important for ubiquitylation, raise a number of issues that, in principle, can be experimentally addressed in Maloideae. Nevertheless, there are also many important differences between the two Rosaceae GSI systems. How such features changed during evolution remains a puzzling issue.

Abstract
We address the problem of predicting the stability of secondary structure motifs of proteins given their linear sequence of residues. Our study is restricted to the prediction of helix structures. We have applied an Inductive Logic Programming (ILP) system to automatically synthesise the predictive rules. ILP systems are well known for being able to induce comprehensible models for data. Furthermore, the models components are definitions provided by a domain expert which makes the model more likely to be helpful in the understanding of the underlying process that produced the data. Our methodology has two stages. First, the system induces a model (set of rules) using just structural information and groupings of the residues to avoid biases by the domain expert. In the second stage, the residues properties are used to make the induced rules Chemically/Biologically appealing. We claim that this methodology is also valuable for general Structure-Activity Relationship (SAR) problems.