Prediction of chromatographic retention by artificial neural networks building a retention index database
Metadata
Show full item recordcomunitat-uji-handle:10234/158176
comunitat-uji-handle2:10234/71345
comunitat-uji-handle3:10234/73147
comunitat-uji-handle4:
TFG-TFMThis resource is restricted
Metadata
Title
Prediction of chromatographic retention by artificial neural networks building a retention index databaseAuthor (s)
Tutor/Supervisor
Pitarch Arquimbau, María ElenaTutor/Supervisor; University.Department
Universitat Jaume I. Departament de Química Física i AnalíticaDate
2016-06-28Publisher
Universitat Jaume IAbstract
Kovats retention index became a tool of paramount importance for the comparison between gas chromatography systems. However, a robust methodology comparable with the Kovats index in relation to liquid chromatographic ... [+]
Kovats retention index became a tool of paramount importance for the comparison between gas chromatography systems. However, a robust methodology comparable with the Kovats index in relation to liquid chromatographic retention time does currently not exist. Furthermore, Artificial Neural Networks (ANNs) have experienced an extraordinary growth during last decade for chromatographic retention prediction. These approaches, however, still have many disadvantages, being the most important of them the system dependency. The prediction can only be applied to a chromatographic system with the conditions used in the construction of the predictive ANN (same chromatographic column, mobile phase and gradient).
In this work, a retention time index strategy has been developed for its application in positive ionisation mode non-target screening analyses based on liquid chromatography coupled to high resolution mass spectrometry allowing an additional identification parameter especially for compounds of which reference standards are not available. A set of 12 isotopically labelled reference standards was applied for the interpolation of retention time indices. Moreover, a mixture of 46 reference standards was used for emulate unknown compounds; and their retention time indices showed a deviation in matrix versus solvent less than 5 % in approximately 94 % of cases. However, its application with a different chromatographic column reduced the percentage of success to 60 % of compounds having deviation below 5%.
Finally, although this investigation is still ongoing, it has been demonstrated the applicability of retention time index in matrix for the correction of retention time shifting and, therefore, avoiding the reporting of false negatives. The extrapolation of the strategy for different chromatographic columns, even working in the 60% of the cases analysed, should be improved. [-]
Subject
Màster Universitari en Tècniques Cromatogràfiques Aplicades | Máster Universitario en Técnicas Cromatográficas Aplicadas | University Master's Degree in Applied Chromatographic Techniques | retention time index | prediction | artificial neural networks (ANNs) | liquid chromatography | isotopically labelled analytical standards
Description
Treball Final del Màster Universitari en Tècniques Cromatogràfiques Aplicades (Pla de 2013). Codi: SIY009. Curs acadèmic 2015-2016
Type
info:eu-repo/semantics/masterThesisRights
http://rightsstatements.org/vocab/CNE/1.0/
info:eu-repo/semantics/restrictedAccess
info:eu-repo/semantics/restrictedAccess