Este trabalho objetivou investigar a existência de polimorfismos de seqüência nucleotídica (SNPs "single nucleotide polymorphisms") em fragmentos genômicos de cana-de-açúcar homólogos a genes de resistência. Para tal, iniciadores sintéticos homólogos às extremidades de uma seqüência expressa identificada (ESTs "expressed sequence tag") como homóloga ao gene Xa1 de arroz (EST SCCCCL3080E03.g) e outra ao gene Rp1-D de milho (EST SCCCCL3080B03.g) foram utilizados para amplificar, em duplicata, fragmentos genômicos de variedades de cana-de-açúcar. Em seguida, os fragmentos foram clonados e seqüenciados. No caso do primeiro EST, foram obtidas seqüências de 119 insertos de uma variedade resistente (SP804966) e 156 de uma variedade suscetível (SP80180) a Xanthomonas albilineans. No segundo caso, foram obtidas seqüências de 167 insertos de uma terceira variedade resistente (R570) e 135 de outra suscetível (SP811763) a Puccinia melanocephala. Para cada EST considerado, as seqüências foram comparadas por meio do programa DNA Sequencher 3.0 (Gene Codes Corporation, Ann Arbor, MI). Nesta análise, somente foram consideradas seqüências reproduzíveis, isto é, que ocorreram nas duas repetições. No caso de ambos ESTs, a comparação de seqüências entre variedades possibilitou a identificação de quatro a seis fragmentos distintos. A comparação entre variedades, por sua vez, revelou a existência de até duas seqüências comuns. Presume-se que estas seqüências reproduzíveis e confirmadas por meio de digestões de clones representativos de cada uma com enzimas de restrição, correspondam a seqüências alélicas. Iniciadores foram sintetizados com base em polimorfismos encontrados entre os diferentes alelos. No entanto, tentativas de amplificar alelos específicos por meio destes iniciadores se revelaram infrutíferas. Tal insucesso deve-se possivelmente à condição polialélica da cana-de-açúcar, que faz com que polimorfismos entre dois alelos sejam compensados por monomorfismos entre estes e os demais alelos. Assim, a utilização de SNPs como marcadores moleculares baseados em PCR em cana-de- açúcar mostra-se mais complexa quando comparada a plantas diplóides. / The objective of this work was to investigate the existence of single nucleotide polymorphisms (SNPs) in genomic fragments homologous to resistance genes in sugarcane. For this purpose, primers were designed based on the sequence of the extremities of an identified expressed sequence tag (ESTs) similar to the Xa1 gene of rice (EST SCCCCL3080E03.g) and another to the maize Rp1-D gene (EST SCCCCL3080B03.g). These primers were used to amplify genomic fragments of sugarcane varieties in duplicate. The fragments were then cloned and sequenced. In the case of first EST, sequences of 119 inserts from a resistant variety (SP804966) and 156 from a susceptible variety (SP80180) to Xanthomonas albilineans were analyzed. In the case of the second EST, sequences of 167 inserts from a third variety (R570) and 135 of a fourth one (SP811763), resistant and susceptible respectively to Puccinia melanocephala, were analyzed. Sequences were compared using the program DNA Sequencher 3.0 (Gene Codes Corporation, Ann Arbor, MI). In this analysis, only reproducible sequences were considered, that is, sequences that occurred in both replicates. Four to six different sequences were identified within varieties whereas comparisons among varieties revealed the existence of up to two sequences in common. These reproducible sequences, which were further confirmed through digestion of representative clones with appropriate restriction enzymes, could correspond to allelic sequences. Primers were designed based on the SNPs detected among these so-called alleles. However, attempts to amplify specific alleles using these primers were unsuccessful. This could be due to the polyallelic condition of sugarcane in which polymorphisms between two alleles could be compensated by monomorphisms between these and other alleles. Thus the use of SNPs as PCR-based molecular markers is not as straightforward in sugarcane as in diploid species.

La forte hausse des rendements agricoles, observée tout au long des 150 dernières années, est, pour une large part, due à l’amélioration des variétés, résultant elle-même essentiellement de processus de recherche et développement. L’optimalité sociale de l’effort de recherche entrepris par les firmes de ce secteur, ainsi que des institutions encadrant l’activité de recherche, constituent un sujet important pour les politiques publiques. L’objet de cette thèse est de contribuer à éclairer cette question, en s’efforçant de tenir compte de l’impact sur l’environnement de l’innovation en matière de variétés agricoles. Nous examinons ce sujet à travers trois prismes différents : celui de l’innovation en tant que telle, celui du processus de recherche, et celui du cadre institutionnel offert aux entreprises de recherche. Nous montrons que la prise en compte des effets environnementaux de la recherche en modifie les optima, et devrait donc conduire à une adaptation du cadre incitatif et règlementaire. / The sharp increase in agricultural yields in the past 150 years owes a lot to the improvement of plant varieties, which, to large extent, is the result of the research and development process. Whether the research and development effort undertaken by firms operating in this sector and institutions regulating research are socially optimal is an important question for public policies. This thesis aims to contribute to tackling this issue, and its main contribution is to endeavor to account for the impact of crop innovation on the environment. We address the question through three different perspectives: innovation per se, the research and development process, and the institutional framework available to research firms. We show that environmental externalities of research significantly modify social optima.