Research On The Application Of Advanced Calibration Models And Pattern Recognition Methods To The Analysis Of Complex Systems

Real-world samples are generally quite complex with matrix effects and background interferences which might have severe influence on the accuracy of the quantitative results of spectrometries and spectrometry-based chemo-bio sensors.Advanced chemometric data analysis methods can effectively mitigate all sorts of interferences in complex spectrometric measurements,and hence improve the accuracy of the quantitative results of spectrometries and chemo-bio sensors for complex real-world samples.In this thesis,attempts have been made to study in detail the feasibility of combining some of the advance calibration models developed by our research group（such as the quantitative fluorescence model for generalized ratiometric probes,the generalized standard addition multivariate calibration strategy based on probe technology）with chemo-bio sensors for accurate quantification of complex real-world samples,and the application of pattern recognition methods to the geographical origin traceability of traditional Chinese medicinal materials.The main research contents are as follows:In Chapter 2,a DNA probe（5’-AGGGTTAGGGTTAGGGTTAGGGAAA/i2-Amp/AAAA-PO4-3’）containing a G4 sequence in its 5’end,a 2-aminopyridine group modified on the fifth nucleic acid base from its 3’end and a phosphate group at its 3’end was designed and used to construct a ratiometric fluorescence sensor for the detection of alkaline phosphatase（ALP）.The proposed ratiometric fluorescence sensor in combination with the quantitative fluorescence model for generalized ratiometric probes(QFM_GRP)was applied to the quantification of ALP in serum samples.Experimental results demonstrated that QFM_GRP could effectively alleviate the detrimental influence of matrix effects and background interferences in serum samples on the quantitative results of the proposed sensor,and helped realize successful accurate quantification of ALP in real-world serum samples with recovery rates ranging from 92.7%to 107%.In Chapter 3,a fluorescence sensor for the detection of formamidopyrimidine DNA glycosylase（Fpg）was designed by integrating the exonucleaseⅢ-assisted signal amplification strategy with a DNA probe（t-DNA）containing an 8-oxyguanine and its complimentary DNA strand（c-DNA）.The proposed sensor was integrated with the generalized standard addition multivariate calibration strategy based on probe technology(GMSA_probe),and achieved accurate quantitative results for Fpg in serum samples with recovery rates varying from 88.1%to 100.6%.In Chapter 4,plasma inductively coupled mass spectrometry（ICP-MS）integrated with pattern recognition methods was applied to the geographical origin traceability of traditional Chinese medicine lily.Lily samples were first processed by ultrasonic assisted acid hydrolysis.Their ICP-MS measurements were directly recorded,and then analyzed by pattern recognition methods.Experimental results revealed that three pattern recognition methods（i.e.,k-nearest neighbor,principal component discriminant analysis and partial least squares discriminant analysis）could accurately discriminate the geographical origin of lily based on their ICP-MS measurements.The correct classification rates were as high as 100%.