No description available.

Made available in DSpace on 2014-06-11T19:24:38Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-08-31Bitstream added on 2014-06-13T20:52:15Z : No. of bitstreams: 1 sartori_aac_me_botfca.pdf: 1545646 bytes, checksum: 2a12c4231007bbdc6a9decf7078f30c2 (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / A Bacia do Rio Pardo representa muito bem a situação de grande parte do território, onde a fragmentação florestal é um dos resultados do processo desordenado de uso e ocupação do solo. O objetivo do presente trabalho foi definir as áreas prioritárias para favorecer a conectividade entre os fragmentos florestais, visando à restauração florestal. Utilizou-se a abordagem multicriterial, em ambiente SIG, com Método da Combinação Linear Ponderada, para combinação dos fatores (características da paisagem) importantes ao objetivo do trabalho. Para estabelecer os fatores, seus respectivos pesos de compensação (expressam a ordem de importância dos fatores no processo de decisão) e restrições (categorias restritivas das características) empregaram-se a Técnica Participatória e o Processo Hierárquico Analítico, no próprio SIG. A princípio os fatores selecionados foram: proximidade entre fragmentos de maior área nuclear; proximidade à cobertura florestal; proximidade à rede hidrográfica; distância aos centros urbanos; erodibilidade; e declividade. Para obtenção desses fatores foram necessários os planos de informação: uso e cobertura do solo; fragmentos de floresta; área nuclear dos fragmentos de floresta; rede hidrográfica; centros urbanos, declividade; e erodibilidade do solo. O resultado obtido foi um mapa de áreas prioritárias, com graus de prioridade representados de forma contínua no espaço de 255 níveis. Para facilitar a visualização e também a análise dos resultados, o mapa foi reclassificado, de maneira a apresentar apenas cinco graus de prioridade. A partir dos resultados obtidos, pode-se concluir que: a Avaliação Multicritérios, em ambiente SIG, é adequada ao mapeamento de áreas prioritárias à restauração florestal em bacias hidrográficas, o método da Combinação Linear Ponderada é flexível, fácil de ser implementado, e permite... / The Pardo River Basin provides a good representation of the situation of the region, where forest fragmentation is a result of the haphazard process of the land use and occupation. The aim of this study was to define the priority areas to facilitate connectivity between forest fragments in order to forest restoration. Multicriteria approach with Weighted Linear Combination was used for combination of factors (landscape features). To establish the factors, their weights compensation (expressing the relative importance of factors in decision making) and the restrictions (restrictive category of characteristics) it was used to Participatory Technical and Analytic Hierarchy Process. The selected factors were: proximity of fragments of larger nuclear area, proximity to forest area, proximity to water, distance to urban centers; soil erodibility, and slope. The result obtained was a map of priority areas, with levels of priority represented on a continuous way, from 0 to 255. To facilitate the visualization and the analysis of results, the map was reclassified, with only five levels of priority. It was concluded, from the results, that the Multicriteria Evaluation ,in a Geographical Information Systems (GIS), is appropriate to the mapping of priority areas for forest restoration in river basins, the method of Weighted Linear Combination is flexible, easy to be implement, and allows the interaction of knowledge (researchers, analysts) and landscape features, determining the robustness of the solution for decision making; multicriteria analysis, based on Analytical Hierarchy Process (AHP), is essential in evaluating the factors, it uses a pair wise comparison between factors to determine the relative importance of each one, reducing subjectivity and subsidizing the creation of better scenarios for the decision maker; and the map of adequacy, for forest recovery River Basin Pardo-SP... (Complete abstract click electronic access below)

Orientador: Célia Regina Lopes Zimback / Banca: Carlos Alberto Vettorazzi / Banca: Sergio Lázaro de Lima / Resumo: A Bacia do Rio Pardo representa muito bem a situação de grande parte do território, onde a fragmentação florestal é um dos resultados do processo desordenado de uso e ocupação do solo. O objetivo do presente trabalho foi definir as áreas prioritárias para favorecer a conectividade entre os fragmentos florestais, visando à restauração florestal. Utilizou-se a abordagem multicriterial, em ambiente SIG, com Método da Combinação Linear Ponderada, para combinação dos fatores (características da paisagem) importantes ao objetivo do trabalho. Para estabelecer os fatores, seus respectivos pesos de compensação (expressam a ordem de importância dos fatores no processo de decisão) e restrições (categorias restritivas das características) empregaram-se a Técnica Participatória e o Processo Hierárquico Analítico, no próprio SIG. A princípio os fatores selecionados foram: proximidade entre fragmentos de maior área nuclear; proximidade à cobertura florestal; proximidade à rede hidrográfica; distância aos centros urbanos; erodibilidade; e declividade. Para obtenção desses fatores foram necessários os planos de informação: uso e cobertura do solo; fragmentos de floresta; área nuclear dos fragmentos de floresta; rede hidrográfica; centros urbanos, declividade; e erodibilidade do solo. O resultado obtido foi um mapa de áreas prioritárias, com graus de prioridade representados de forma contínua no espaço de 255 níveis. Para facilitar a visualização e também a análise dos resultados, o mapa foi reclassificado, de maneira a apresentar apenas cinco graus de prioridade. A partir dos resultados obtidos, pode-se concluir que: a Avaliação Multicritérios, em ambiente SIG, é adequada ao mapeamento de áreas prioritárias à restauração florestal em bacias hidrográficas, o método da Combinação Linear Ponderada é flexível, fácil de ser implementado, e permite... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The Pardo River Basin provides a good representation of the situation of the region, where forest fragmentation is a result of the haphazard process of the land use and occupation. The aim of this study was to define the priority areas to facilitate connectivity between forest fragments in order to forest restoration. Multicriteria approach with Weighted Linear Combination was used for combination of factors (landscape features). To establish the factors, their weights compensation (expressing the relative importance of factors in decision making) and the restrictions (restrictive category of characteristics) it was used to Participatory Technical and Analytic Hierarchy Process. The selected factors were: proximity of fragments of larger nuclear area, proximity to forest area, proximity to water, distance to urban centers; soil erodibility, and slope. The result obtained was a map of priority areas, with levels of priority represented on a continuous way, from 0 to 255. To facilitate the visualization and the analysis of results, the map was reclassified, with only five levels of priority. It was concluded, from the results, that the Multicriteria Evaluation ,in a Geographical Information Systems (GIS), is appropriate to the mapping of priority areas for forest restoration in river basins, the method of Weighted Linear Combination is flexible, easy to be implement, and allows the interaction of knowledge (researchers, analysts) and landscape features, determining the robustness of the solution for decision making; multicriteria analysis, based on Analytical Hierarchy Process (AHP), is essential in evaluating the factors, it uses a pair wise comparison between factors to determine the relative importance of each one, reducing subjectivity and subsidizing the creation of better scenarios for the decision maker; and the map of adequacy, for forest recovery River Basin Pardo-SP... (Complete abstract click electronic access below) / Mestre

Multifunktionella aspekter och flerdimensionella problem öppnar upp för användandet av geografiska informationssystem (GIS) vid beslutsfattande. Ett användningsområde i GIS är möjligheten att ta fram lämpliga områden, det innebär områden där flera lämpliga aspekter stämmer överens med varandra. Denna typ av lämplighetsanalys påverkas i hög grad av författaren samt beroende på analysmetod vilket i sin tur påverkar beslutsfattande. Denna studie syftar till att undersöka hur de val författaren gör påverkar metoden samt vilka konsekvenser det får för beslutsprocessen. Utifrån tidigare forskning har två analysmetoder identifierats, Analytic Hierarchy Process (AHP) och Weighted Linear Combination (WLC). För att kunna jämföra de två metoderna tar studien stöd i en exempelstudie från Linköpings tätort. Utifrån denna studie går det att se att författaren påverkar resultatet av kriterier och viktning men där det skiljer sig olika mycket mellan de två analysmetoderna AHP och WLC. AHP och WLC har både styrkor och svagheter som gör dem lämpliga för denna sorts studie. / Multifunctional aspects and multidimensional problems open up the use of geographic information systems (GIS) in decision-making. One area of ​​use in GIS is the site selection, the site selection combines areas with several suitable aspects. This type of suitability analysis is greatly influenced by the author and depending on the analysis method, which in turn affects decision-making. This study aims to investigate how the choices the author makes affect the method and what consequences it has for the decision-making process. Based on previous research, two analysis methods have been identified, Analytic Hierarchy Process (AHP) and Weighted Linear Combination (WLC). In order to compare the two methods, the study is supported by an example study from Linköping City. Based on this study, it can be seen that the author influences the results of criteria and weighting. However the two analysis methods AHP and WLC affect the study in different ways. AHP and WLC both have strengths and weaknesses that make them suitable for site selection studies.

[Abstract]: The disposal of animal waste has become a problem in many parts of the world due to the rapid growth in the number and the size of intensive animal industries. Safe waste disposal sites are rarely available and the relocation and/or treatment of animal waste is seldom economically viable. The reuse of animal waste for energy recovery and re-feeding is also not popular. Animal waste is a valuable source of plant nutrients and a very good soil conditioner, and has been commonly applied as fertiliser to agricultural fields. However, due to the increasing oversupply of animal waste in recent years, it has often been applied in excess to the agricultural fields. Excessive application of animal waste, without due consideration of its implications, is a serious concern. The run-off and leaching losses of nutrients from the fields fertilised with animal waste have contributed significantly to the eutrophication and toxic blue-green algae blooms in surface water systems and nitrification of ground water systems. It has also led to nutrient imbalances in the soils and odour pollution to the surrounding communities. The animal waste, which is a valuable source of plant nutrients, has thus become both an economic and environmental burden, and there is a need to develop a strategy for its sensible use as a fertiliser in agricultural fields. Sensible use of animal waste involves the consideration of all the agricultural, environmental, social, and economical limitations. A rational method of achieving this is to restrict the use of animal waste to sites suitable for such uses, identify areas where it can be relocated and applied economically, limit the application rates to a safe level, and observe appropriate manure management practices. This study addressed each of these components by developing a comprehensive manure application plan (MAP) for the site-specific use of animal waste as fertiliser in agricultural fields. Various geographic information systems (GIS) based techniques, including a weighted linear combination model and map algebra based cartographic modelling, were employed to achieve the goal. The appropriateness of the existing techniques and procedures were evaluated and modified to meet the current input requirements. New methods of analysis were devised as necessary. The Westbrook sub-catchment of the Condamine River catchment in south-east Queensland was selected as the study area. The sub-catchment covers 24,903 hectares and contains 39 intensive animal industries. The catchment is also a part of the Murray-Darling Basin, which has been suffering from toxic blue green algae blooms recurrently since 1991. This study identified that only about one-fifth of the sub-catchment area is suitable for animal waste application. Depending on the method of site suitability analysis and the number of input factors used the suitable area ranged between 16 and 22 percent. This comparatively small area is mainly due to the presence of a large proportion of non-agricultural areas in the sub-catchment. The suitable areas were also found to have various degrees of suitability for waste application. However, the degree of site suitability was affected by the number of input factors used in the analysis, the weighting of the factors, and the method of factor attribute standardisation. Conventional methods of weighting input factors were found to be cumbersome and not particularly suitable. Hence, this study developed a new ‘objective oriented comparison’ method of factor weighting. Standardisation of input factors using a continuous, rather than discrete, classification (ie fuzzy set) method was found to be more consistent in degree of suitability determination. The discrete classification of factor attributes into classes of different numbers and sizes, and the weighting of classes to a sum of one, were identified as a limitation in using this standardisation method. A new ‘weight adjustment’ method was devised and demonstrated to reduce factor-weighting biases. The suitable sites, degree of site suitability, and other relevant spatial and non-spatial information were processed within a GIS framework to develop a comprehensive manure application plan. The inherently high presence of available phosphorus in the soils of the study area was recognised and the P2O5 content in the manure was used as the basis for determining manure application rates. A complimentary nitrogen supply map was also generated. Manure management practices applicable to the areas with a lower degree of suitability were also suggested.

Os movimentos de massa são fenômenos naturais que fazem parte dos processos de dinâmica superficial que modelam a superfície terrestre. Quando ocorrem em áreas ocupadas, podem provocar perdas econômicas, impactos sociais e, em casos extremos, perda de vidas humanas. O estabelecimento de medidas preventivas para mitigar estes danos demanda a identificação das áreas mais propensas a esses fenômenos. Partindo deste contexto, o objetivo desta dissertação foi desenvolver uma proposta metodológica para mapear a suscetibilidade a escorregamentos no município de Porto Alegre-RS, com o emprego de ferramentas de geoprocessamento. As variáveis utilizadas foram: declividade, geologia, distância de lineamentos, acúmulo de fluxo, uso do solo e cobertura vegetal, as quais foram processadas no software Idrisi. Após a padronização das variáveis por meio da aplicação de funções fuzzy e pela atribuição direta de notas, definiu-se a importância de cada uma na predisposição do terreno a escorregamentos com base na consulta a especialistas, com o apoio da técnica denominada processo analítico hierárquico. A integração das variáveis foi realizada por meio de combinação linear ponderada, estabelecendo-se dois cenários de suscetibilidade a escorregamentos: o primeiro com a suscetibilidade natural e o segundo com a suscetibilidade induzida. Os resultados indicam que as áreas mais suscetíveis a escorregamentos situam-se, principalmente, nos morros da Crista de Porto Alegre e nos morros isolados no sul do município. O restante do município apresenta, de maneira geral, baixa suscetibilidade. A validação dos resultados obteve uma exatidão geral de 95% e 94% respectivamente para os Cenários 1 e 2. As áreas abaixo da curva ROC (Receiver Operating Characteristic) foram respectivamente de 0,960 e 0,966, o que indica que as variáveis e os pesos foram adequados e que ambos os cenários são robustos. Assim, constata-se que as ferramentas de geoprocessamento são úteis no mapeamento da suscetibilidade a escorregamentos, pois tornam ágil o mapeamento e sua atualização, e dão mais clareza e transparência ao processo de avaliação da suscetibilidade. Além disso, o método proposto permite reduzir os esforços nas etapas iniciais da gestão de riscos, possibilitando ao gestor priorizar os mapeamentos mais detalhados nas áreas com maior suscetibilidade. Isso demonstra que as rotinas de apoio à decisão, como as utilizadas neste trabalho, podem ser incorporadas em estudos de suscetibilidade, provendo importantes informações para o planejamento urbano dos municípios. / Mass movements are natural phenomena that are part of the surface dynamic processes that shape the Earth’s surface. When they occur in urban areas, can cause economic losses, social impacts and, in extreme cases, losses of human lives. The implementation of preventive measures to mitigate those damages requires the identification of the areas prone to these events. Therefore, the aim of this dissertation was to develop a methodological approach to map landslide susceptibility at the municipality of Porto Alegre-RS, with the use of geoprocessing tools. The variables used were: slope, geology, distance from lineaments, flow accumulation and land use, which were processed in the Idrisi software. After the standardization of the variables through the application of fuzzy functions and the direct assignment of notes, the importance of each one on the predisposition of landslides was defined, based on the opinion of experts, with the support of the technique entitled analytic hierarchy process (AHP). The integration of the variables was based on weighted linear combination, resulting in two scenarios: the first with the natural susceptibility and the second with the induced susceptibility. The results indicate that the most susceptible areas to landslide are located in the hills of the Crista de Porto Alegre and in the isolated hills located at the south of the municipality. The remaining area has, in general, low susceptibility. The validation of the results obtained an overall accuracy of 95% and 94%, respectively for Scenarios 1 and 2. The areas under the ROC (Receiver Operating Characteristic) curve were respectively 0.960 and 0.966, indicating that the variables and weights were adequate and both scenarios are robust. Thus, it is noted that geoprocessing tools are useful to map the landslide susceptibility, since they make the mapping process and its updates agile, and give more clarity and transparency to the susceptibility assessment process. In addition, the proposed method reduces efforts in the early stages of risk management, enabling the manager to prioritize the most susceptible areas for detailed mappings. This demonstrates that decision support routines, such as those used in this dissertation, can be incorporated into susceptibility studies, providing important information for the urban planning.

Os movimentos de massa são fenômenos naturais que fazem parte dos processos de dinâmica superficial que modelam a superfície terrestre. Quando ocorrem em áreas ocupadas, podem provocar perdas econômicas, impactos sociais e, em casos extremos, perda de vidas humanas. O estabelecimento de medidas preventivas para mitigar estes danos demanda a identificação das áreas mais propensas a esses fenômenos. Partindo deste contexto, o objetivo desta dissertação foi desenvolver uma proposta metodológica para mapear a suscetibilidade a escorregamentos no município de Porto Alegre-RS, com o emprego de ferramentas de geoprocessamento. As variáveis utilizadas foram: declividade, geologia, distância de lineamentos, acúmulo de fluxo, uso do solo e cobertura vegetal, as quais foram processadas no software Idrisi. Após a padronização das variáveis por meio da aplicação de funções fuzzy e pela atribuição direta de notas, definiu-se a importância de cada uma na predisposição do terreno a escorregamentos com base na consulta a especialistas, com o apoio da técnica denominada processo analítico hierárquico. A integração das variáveis foi realizada por meio de combinação linear ponderada, estabelecendo-se dois cenários de suscetibilidade a escorregamentos: o primeiro com a suscetibilidade natural e o segundo com a suscetibilidade induzida. Os resultados indicam que as áreas mais suscetíveis a escorregamentos situam-se, principalmente, nos morros da Crista de Porto Alegre e nos morros isolados no sul do município. O restante do município apresenta, de maneira geral, baixa suscetibilidade. A validação dos resultados obteve uma exatidão geral de 95% e 94% respectivamente para os Cenários 1 e 2. As áreas abaixo da curva ROC (Receiver Operating Characteristic) foram respectivamente de 0,960 e 0,966, o que indica que as variáveis e os pesos foram adequados e que ambos os cenários são robustos. Assim, constata-se que as ferramentas de geoprocessamento são úteis no mapeamento da suscetibilidade a escorregamentos, pois tornam ágil o mapeamento e sua atualização, e dão mais clareza e transparência ao processo de avaliação da suscetibilidade. Além disso, o método proposto permite reduzir os esforços nas etapas iniciais da gestão de riscos, possibilitando ao gestor priorizar os mapeamentos mais detalhados nas áreas com maior suscetibilidade. Isso demonstra que as rotinas de apoio à decisão, como as utilizadas neste trabalho, podem ser incorporadas em estudos de suscetibilidade, provendo importantes informações para o planejamento urbano dos municípios. / Mass movements are natural phenomena that are part of the surface dynamic processes that shape the Earth’s surface. When they occur in urban areas, can cause economic losses, social impacts and, in extreme cases, losses of human lives. The implementation of preventive measures to mitigate those damages requires the identification of the areas prone to these events. Therefore, the aim of this dissertation was to develop a methodological approach to map landslide susceptibility at the municipality of Porto Alegre-RS, with the use of geoprocessing tools. The variables used were: slope, geology, distance from lineaments, flow accumulation and land use, which were processed in the Idrisi software. After the standardization of the variables through the application of fuzzy functions and the direct assignment of notes, the importance of each one on the predisposition of landslides was defined, based on the opinion of experts, with the support of the technique entitled analytic hierarchy process (AHP). The integration of the variables was based on weighted linear combination, resulting in two scenarios: the first with the natural susceptibility and the second with the induced susceptibility. The results indicate that the most susceptible areas to landslide are located in the hills of the Crista de Porto Alegre and in the isolated hills located at the south of the municipality. The remaining area has, in general, low susceptibility. The validation of the results obtained an overall accuracy of 95% and 94%, respectively for Scenarios 1 and 2. The areas under the ROC (Receiver Operating Characteristic) curve were respectively 0.960 and 0.966, indicating that the variables and weights were adequate and both scenarios are robust. Thus, it is noted that geoprocessing tools are useful to map the landslide susceptibility, since they make the mapping process and its updates agile, and give more clarity and transparency to the susceptibility assessment process. In addition, the proposed method reduces efforts in the early stages of risk management, enabling the manager to prioritize the most susceptible areas for detailed mappings. This demonstrates that decision support routines, such as those used in this dissertation, can be incorporated into susceptibility studies, providing important information for the urban planning.

Os movimentos de massa são fenômenos naturais que fazem parte dos processos de dinâmica superficial que modelam a superfície terrestre. Quando ocorrem em áreas ocupadas, podem provocar perdas econômicas, impactos sociais e, em casos extremos, perda de vidas humanas. O estabelecimento de medidas preventivas para mitigar estes danos demanda a identificação das áreas mais propensas a esses fenômenos. Partindo deste contexto, o objetivo desta dissertação foi desenvolver uma proposta metodológica para mapear a suscetibilidade a escorregamentos no município de Porto Alegre-RS, com o emprego de ferramentas de geoprocessamento. As variáveis utilizadas foram: declividade, geologia, distância de lineamentos, acúmulo de fluxo, uso do solo e cobertura vegetal, as quais foram processadas no software Idrisi. Após a padronização das variáveis por meio da aplicação de funções fuzzy e pela atribuição direta de notas, definiu-se a importância de cada uma na predisposição do terreno a escorregamentos com base na consulta a especialistas, com o apoio da técnica denominada processo analítico hierárquico. A integração das variáveis foi realizada por meio de combinação linear ponderada, estabelecendo-se dois cenários de suscetibilidade a escorregamentos: o primeiro com a suscetibilidade natural e o segundo com a suscetibilidade induzida. Os resultados indicam que as áreas mais suscetíveis a escorregamentos situam-se, principalmente, nos morros da Crista de Porto Alegre e nos morros isolados no sul do município. O restante do município apresenta, de maneira geral, baixa suscetibilidade. A validação dos resultados obteve uma exatidão geral de 95% e 94% respectivamente para os Cenários 1 e 2. As áreas abaixo da curva ROC (Receiver Operating Characteristic) foram respectivamente de 0,960 e 0,966, o que indica que as variáveis e os pesos foram adequados e que ambos os cenários são robustos. Assim, constata-se que as ferramentas de geoprocessamento são úteis no mapeamento da suscetibilidade a escorregamentos, pois tornam ágil o mapeamento e sua atualização, e dão mais clareza e transparência ao processo de avaliação da suscetibilidade. Além disso, o método proposto permite reduzir os esforços nas etapas iniciais da gestão de riscos, possibilitando ao gestor priorizar os mapeamentos mais detalhados nas áreas com maior suscetibilidade. Isso demonstra que as rotinas de apoio à decisão, como as utilizadas neste trabalho, podem ser incorporadas em estudos de suscetibilidade, provendo importantes informações para o planejamento urbano dos municípios. / Mass movements are natural phenomena that are part of the surface dynamic processes that shape the Earth’s surface. When they occur in urban areas, can cause economic losses, social impacts and, in extreme cases, losses of human lives. The implementation of preventive measures to mitigate those damages requires the identification of the areas prone to these events. Therefore, the aim of this dissertation was to develop a methodological approach to map landslide susceptibility at the municipality of Porto Alegre-RS, with the use of geoprocessing tools. The variables used were: slope, geology, distance from lineaments, flow accumulation and land use, which were processed in the Idrisi software. After the standardization of the variables through the application of fuzzy functions and the direct assignment of notes, the importance of each one on the predisposition of landslides was defined, based on the opinion of experts, with the support of the technique entitled analytic hierarchy process (AHP). The integration of the variables was based on weighted linear combination, resulting in two scenarios: the first with the natural susceptibility and the second with the induced susceptibility. The results indicate that the most susceptible areas to landslide are located in the hills of the Crista de Porto Alegre and in the isolated hills located at the south of the municipality. The remaining area has, in general, low susceptibility. The validation of the results obtained an overall accuracy of 95% and 94%, respectively for Scenarios 1 and 2. The areas under the ROC (Receiver Operating Characteristic) curve were respectively 0.960 and 0.966, indicating that the variables and weights were adequate and both scenarios are robust. Thus, it is noted that geoprocessing tools are useful to map the landslide susceptibility, since they make the mapping process and its updates agile, and give more clarity and transparency to the susceptibility assessment process. In addition, the proposed method reduces efforts in the early stages of risk management, enabling the manager to prioritize the most susceptible areas for detailed mappings. This demonstrates that decision support routines, such as those used in this dissertation, can be incorporated into susceptibility studies, providing important information for the urban planning.

Threading is a protein structure prediction method that uses a library of template protein structures in the following steps: first the target sequence is matched to the template library and the best template structure is selected, secondly the predicted target structure of the target sequence is modeled by this selected template structure. The deceleration of new folds which are added to the protein data bank promises completion of the template structure library. This thesis uses a new set of template-specific weights to improve the energy function for sequence-to-structure alignment in the template selection step of the threading process. The weights are estimated using least squares methods with the quality of the modelling step in the threading process as the label. These new weights show an average 12.74% improvement in estimating the label. Further family analysis show a correlation between the performance of the new weights to the number of seeds in pFam.

Threading is a protein structure prediction method that uses a library of template protein structures in the following steps: first the target sequence is matched to the template library and the best template structure is selected, secondly the predicted target structure of the target sequence is modeled by this selected template structure. The deceleration of new folds which are added to the protein data bank promises completion of the template structure library. This thesis uses a new set of template-specific weights to improve the energy function for sequence-to-structure alignment in the template selection step of the threading process. The weights are estimated using least squares methods with the quality of the modelling step in the threading process as the label. These new weights show an average 12.74% improvement in estimating the label. Further family analysis show a correlation between the performance of the new weights to the number of seeds in pFam.