The Determination Of Biomass,Glycerol And Methanol In Fermentation Process Based On Near Infrared Spectroscopy And Aquaphotomics

In recent years,due to the rapid development of the biopharmaceutical industry,antibody drugs as its pillar industry are getting more and more attention.Because its quality and yield are affected by some key parameters in the fermentation process,which needs to be measured in the production process.In the fermentation process of pichia pastoris,the increase of biomass and the consumption of glycerol in the first stage are crucial to the selection of induction time.The accurate selection of logarithmic growth period can save more energy and time in the production,which is conducive to obtaining sufficient number of yeast cells with strong metabolism.In the second stage,methanol,as the promoter,is of vital importance.Too low concentration methanol will lead to cell lysis and the difficult polymerization of secreted proteins.On the contrary,high concentration will inhibit cell growth and protein expression,so the above three parameters need to be measured and controlled in the production process.However,two difficulties are needed to overcome in the process of parameter determination.Firstly,the accuracy of the spectra is needed to be improved because the complexity of the fermented liquid and the disturbance of the bacteria,so subtraction methods were introduced to this study and previous research has shown the advantages of subtraction methods,our study compared three methods to improve the quality of the spectrum to the greatest extent(subtract directly,subtract in turn and least-area method).Secondly,the content of glycerol and methanol is relatively low,and near-infrared spectroscopy(NIRS)is a secondary method.Hence,aquaphotomics was introduced in this study to solve the problem,which can use the influence of disturbance factors on the water spectrum to achieve the monitoring of other substances,and this technology has been successfully applied to monitor low concentration of proteins,sugars,etc.This provides a new possibility for the accurate monitoring of glycerol and methanol content.Therefore,this study inspect the feasibility of subtraction methods and aquaphotomics using simple systems,and then these combined with NIRS technology were introduced into the measurement of biomass,glycerin and methanol in the fermentation process of pichia pastoris to further explore the feasibility of NIRS technology combined with subtraction methods and aquaphotomics in the fermentation precess.The details of this paper are listed as below:(1)Reserch on three different subtraction methods.In this chapter,Antaris ? Fourier transform near-infrared spectrometer was used to explore the advantages of three kinds of subtraction methods(subtract directly,subtract in turn and least-area method)in the simple methanol-water solution system,so as to improve the spectral quality to the maximum extent.Firstly,methanol-water mixture was prepared with methanol content from 0.1%to 2.5%(v/v)with increments of 0.1%.Four independent replicates were prepared,two of the replicates were made with direct dilution,while the other two replicates were prepared with serial dilution,so total number of samples in the experiment was 100.The regression coefficients of the partial least squares regression(PLSR)model were used to investigate the direct and serial dilution methods.The results showed that the two solutions were not significantly different from each other and have no significance influence the accuracy of the model.Cross validation of root mean square error(RMSECV)value was chose as the main criteria,the results showed that the model built with subtraction in turn had the best performance based on direct dilution method,the middle is the least-area method and the worst is the direct subtraction,but still better than the traditional method of averaging three spectra and the RMSECV values were reduced by about half.(2)The feasibility study of aquaphotomics in fermentation process.This study investigated the feasibility of NIRS combined with aquaphotomics for methanol determination in methanol-medium(yeast extract peptone dextrose,YPD,luria-bertani,LB,yeast extract peptone ethanol glycerin,YEPG)system.Firstly,methanol-medium(YPD,LB,YEPG)solutions were prepared with a concentration range of 0.1-2.5%(v/v)with an increasement of 0.1%prepared with direct dilution method.There were 25 samples in each medium.Secondly,12 water matrix coordinates(WAMACS,7149 cm-1,7082 cm-1,6954 cm-1,6940 cm-1,6900 cm-1,6871 cm-1,6836 cm-1,6773 cm-1,6702 cm-1,6667 cm-1,6509 cm-1,6400 cm-1)were selected to reflect the influence of methanol on the water spectrum through aquagrams.The results showed that the absorbance could be affected by the addition of 0.1%methanol.Then,PLSR was chosen to build quantitative models for methanol determination.PLSR was used to establish a regression analysis model for the determination of low methanol content.As an example,the RMSEC,RMSECV,RMSEP,Rc2,Rcv2 and Rp2 values of the optimal methanol-YPD model were 0.014%,0.038%,0.047%,1.000,0.997 and 0.999,respectively.The results of methanol-LB and methanol-YEPG medium were similar to the above results,which proved that NIR could be used for the determination of methanol content and provided a new idea for the determination of low methanol content.(3)Research on the key parameters determination in the fermentation process:biomass,glycerol and methanol.In this study,the above conclusions were used to determine the biomass,glycerol and methanol content in pichia pastoris fermentation process.First,four(first stage)and other five(second stage)parallel tests were conducted at the shaking bottle level,then 100 and other 110 samples were colected from the first and second stages of pichia pastoris fermentation process,respectively,in which the first stage samples contained changes in biomass and glycerol and the second stage samples contained the changes in methanol.The primary data of the three parameters were determined by vis-UV,glycerol assay kit and gas chromatography,respectively.The methods of outlier detection,dataset dividing and spectral preprocessing were optimized before modeling.Finally,the RMSEC,RMSECV,RMSEP,Rc2,Rcv2 and Rp2 values of the best LWR model of biomass were 0.022,0.087,0.116,0.999,0.987 and 0.978,respectively.The values of best LWR model of glycerol were 0.243,1.037,0.765,0.997,0.951 and 0.963,respectively.The values of the best PLSR model of methanol were 0.128,0.141,0.149,0.924,0.907 and 0.852,respectively.These three models can realize the quantitative analysis of biomass,glycerol and methanol in the fermentation process of pichia pastoris rapidly and accuratelyThe innovation points are listed as below(1)In this study,12 WAMACS were selected to reflect the influence of methanol on the water spectrum and the feasibility of using aquaphotomics in the fermentation process was investigated for the first time(2)In this section,NIRS combined with the subtraction method and aquaphotomics was used to monitor the key quality attributes(biomass,glycerol and methanol)in the fermentation process for the first time.