Research On Near Infrared Spectroscopy In The Determination Of Source Plasma Protein

Blood product is defined as the biological products that depend on the biological process, method of separation and purification technology to separate and extract active ingredients, including a variety human plasma protein products. In order to ensure the quality of raw plasma, regulatory authorities have developed some strict standards. The collected plasma not only should be determined protein content, alanine aminotransferase (ALT), HBSAg, syphilis, HIV-1antibody, HIV-2antibody and HCV antibody,but also requires a three-month quarantine period. The chemical composition of the plasma, water accounts for90～92%and other10%is mostly plasma protein, besides, nutrients (nutrients), enzymes (enzymes), hormones (hormones), cholesterol (cholesterol) and other important components. Plasma protein is the general term for a variety of protein, which can be divided into albumin, globulin and fibrinogen categories by salt analysis method. At present, China Pharmacopoeia of single-dose plasma protein content can be used biuret or refractometer determination should not be less than50g/L. Which are based on biuret protein peptid bonds in alkaline solution to form puRle complexes with Cu2+, And its ptotein content is proportional to the depth and color, using standard protein solution as control, using ultraviolet-visible spectrophotometry for test protein The determination of protein content, a single reagent method is simple, but the sensitivity is poor, time-consuming. Refractometer method is a faster way to get protein, but this method and other methods the results have large errors. NIR wavelength range of visible light is in the infrared region and between the United States Materials Testing Association to define the wavelength range780-2526nm. Near-infrared spectroscopy contains a lot of material information, mainly hydrogen groups, such as OH, CH, NH and other groups to absorb the vibration frequency multiplier harmony, a wide range near-infrared spectroscopy, in agricultural products, feed, drugs, petrochemicals other fields have achieved a very good application of the results. In recent years, near-infrared also repeated reports for medical testing, such as successfully applied to brain tissue and blood oxygen test,blood glucose, oxygenated hemoglobin, deoxy hemoglobin and cytochrome oxidase in vivo near-infrared non-destructive testing and so on.Near Infrared Spectroscopy for its rapid, nondestructive and accurate characteristics to be more widely used, this topic aims to establish a rapid and accurate method of plasma protein targets, and through different methods of band selection, spectral pre-method, modeling method optimization model in order to get the best results and prediction model for the future near-infrared spectroscopy in the raw plasma process to lay a foundation application. When the model was successfully established, it can guarantee the health of plasma donor and decrease the time that they stay in the plasma stations, which increase the efficiency. Furthermore, thousands ton of plasma needs to be recheck every year. So, a successful model can provide methods to detect the content of protein in physical production.This subject proved the feasibility of establishing a model about protein content in raw plasma and samples are prepared with different physiological saline. The PLS model has a good prediction capability after investigating different pre-treatment methods and wave selection methods. The PLS model, using SG9-point smoothing and seven principal component scores, the Rcv is0.99965, RMSECV is0.464, R is0.99986, RMSEP is0.324.Then, model of protein concentration in raw plasma was established after comparation of the pretreatments methods, including the first derivative, second derivative, SG smooth, Norris smoothing and the comparation of different wavelength range selection methods, including correlation coefficient method, iPLS method. Finally, model established in6500-5500cm-1has the best accuracy and precision and the repeatability and could meet the actual requirements, the parameters are R=0.99323, RMSEP=3.92, Rcv=0.99214, RMSECV=2.4.