Font Size: a A A

Antitumor Active Ingredients Identification From Curcuma Longa L.(Turmeric) Total Extracts Based On ANNs And SVR

Posted on:2015-06-30Degree:MasterType:Thesis
Country:ChinaCandidate:Z D LiFull Text:PDF
GTID:2284330452958676Subject:Pharmaceutical Engineering
Abstract/Summary:PDF Full Text Request
Turmeric, dried rhizome of Curcuma longa L.(Zingiberaceae), has abundantedible and medicinal value. Its efficacy mainly comes from turmeric essential oil andcurcuminoids. The target of this study was to identify the antitumor compounds fromturmeric total extracts (TTE, including essential oil and curcuminoids), finally helpfulfor the new drug disvery from Curcuma longa L.. The main research was stated asfollows:1. Accomplished6extract methods and6optimization methods for TTE. Throughanalysis the extract yields and composition, finally the extract process was as follows:EtOH ultra-sonication, solid-liquid ratio1:5, twice with each time15min. By thismethod,48batches of TTE were prepared with yields ranged from4.62%to11.22%(g/g, TTE/rhizomes).2. HPLC-MS (finished by cooperator) and GC-MS were used to analyze thecomposition of48batches of TTE.18non-volatile ingredients and19volatileingredients were analyzed qualitatively.22non-volatile ingredients were quantifiedvia relative peak area (RPA)(A/As2, s2: isoquercitrin), meanwhile,21volatileingredients were quantified via RPA (A/As1, s1: n-tridecane). The cytotoxicity of48samples TTE against HeLa cells were detected by an MTT assay and the inhibitionratios ranged from42.4%(sample19) to86.3%(sample43).3. The composition-activity relationship (CAR) model for TTE was constructedthrough artificial neural network (ANNs, including BPNN and GRNN) and supportvector regression (SVR), coupled with parameter optimization methods (GA, PSO,GS and k-fold cross-validation). Through compared errors (RMSE, RSE) andcorrelation coefficient (R), GA-SVR was the super model (RMSE, RSE and R of testdata were0.013,0.021and0.979, respectively), followed by GRNN (RMSE, RSEand R of test data were0.015,0.024and0.970, respectively).4.13ingredients from TTE were identified with marked anticancer activitythrough mean impact value (MIV) combined with GA-SVR and GRNN respectively,which include6volatile ingredients (ar-turmerone,-turmerone, β-turmerone,ar-curcumene, β-sesquiphellandrene and-zingiberene) and7non-volatile ingredients(demethoxycurcumin, curcumin, bisdemethoxycurcumin, cyclocurcumin,demethoxycyclocurcumin and2unknown ingredients). Except two unknowningredients, the remaining11constituents were identified with cytotoxicity by previous researches. The results indicated that the proposed method was credible andreferable for active ingredients identification.
Keywords/Search Tags:turmeric total extracts, GC-MS, ANNs, SVR, active ingredientsidentification
PDF Full Text Request
Related items