核磁共振 波譜分析

1、基于核磁共振的代謝組學(xué)數(shù)據(jù)分析標(biāo)準(zhǔn)化,代謝組學(xué)中的動(dòng)物實(shí)驗(yàn)設(shè)計(jì)生物樣品采集、制備及核磁共振檢測(cè)代謝組學(xué)中的數(shù)據(jù)分析核磁共振譜圖預(yù)處理核磁共振數(shù)據(jù)前處理多變量統(tǒng)計(jì)分析,主要內(nèi)容,代謝組學(xué)中的動(dòng)物實(shí)驗(yàn)設(shè)計(jì),動(dòng)物的選擇不符合3R原則，多多益善 Reduction(減少)，Replacement(替代)， Refinement(優(yōu)化)從微生物學(xué)標(biāo)準(zhǔn)選擇動(dòng)物 普通級(jí)，清潔級(jí)，SPF級(jí)，無(wú)菌動(dòng)物從效果上選擇

2、實(shí)驗(yàn)動(dòng)物 常用的動(dòng)物實(shí)驗(yàn)順序：小鼠、大鼠 (嚙齒目)、兔 (兔形目)、豬、牛 (偶蹄目)、狗、貓 (食肉目) 、獼猴、黑猩猩(靈長(zhǎng)目)等。,人和不同種實(shí)驗(yàn)動(dòng)物血清的NMR譜圖比較,屬(species)，種(strain),人和不同種實(shí)驗(yàn)動(dòng)物尿樣的NMR譜圖比較,人和不同種實(shí)驗(yàn)動(dòng)物的代謝輪廓比較,,D: Dairy Cow; H: Human;M: Mouse; P: Piggy;R: Rat; G: Guinea Pig,M-

3、B: Balb-C Mouse; H: Human;M-K: Kunming Mouse; Ra: Rabbit;R-S: SD Rat; R-W: Wistar Rat,性別(gender),血樣,尿樣,,1,,2,1,,,Zebrafish Livers OPLS,1H NMR (800 MHz) profile of liver from (A) Male and (B) Female Zebrafish,Ong ES et

4、al Mol. Biosystems, 2009;5:288-98.,性別(gender),代謝組學(xué)中的動(dòng)物實(shí)驗(yàn)設(shè)計(jì),年齡(Age range) 年齡不同，生物學(xué)特性各異，一般選性成熟后的青壯年動(dòng)物。 太小的動(dòng)物：生理功能未達(dá)到成年水平。太老的動(dòng)物：各器官老化，代謝功能下降 體重(Weight range) 在正常營(yíng)養(yǎng)狀態(tài)和飼養(yǎng)條件下，體重與年齡有一定的相關(guān)性。,實(shí)驗(yàn)動(dòng)物年齡狀態(tài)的相似性,代謝組學(xué)中的動(dòng)物實(shí)驗(yàn)設(shè)計(jì),飲

5、食(diet control)環(huán)境(environment) 溫度(18-25ºC)，濕度(45-65%)喂養(yǎng)(housing) 12h-12h 光照/黑暗適應(yīng)期(acclimation)……,生物樣本的采集,消毒過(guò)程和麻醉劑的影響,消毒過(guò)程的影響,抗凝劑的影響,生物樣本的采集,注意事項(xiàng)尿樣和糞樣收集 采樣時(shí)間一致(8：00-12：00)，避免晝夜節(jié)律產(chǎn)生的代謝變化。加NaN3防酶、抗菌。血樣收集

6、 取血前禁食12小時(shí)。避免使用酒精消毒。組織收集 盡可能在15分鐘內(nèi)收集完整。,生物樣本的制備,生物組織與細(xì)胞,重水配制的生理鹽水,HR-MAS轉(zhuǎn)子,,,生物樣本的制備,代謝組學(xué)生物樣本的核磁共振檢測(cè),生物樣本的NMR檢測(cè),代謝組中包含的代謝物類(lèi)型,Metabolome = the total metabolite pool,- All low molecular weight (MW < 1000 Da) org

7、anic molecules in a biological sample.,PeptidesOligonucleotidesSugarsNucleosidesOrganic acidsKetonesAldehydesAminesAmino acidsLipidsSteroidsAlkaloidsDrugs (xenobiotics),Metabolic pathway,Metabolites detected

8、by NMR,(+)-(-)-Methylsuccinic Acid2,5-Dihydroxyphenylacetic Acid2-hydroxy-3-methylbutyric acid2-Oxoglutaric acid3-Hydroxy-3-methylglutaric acid3-Indoxyl Sulfate 5-Hydroxyindole-3-acetic AcidAcetamideAceti

9、c AcidAcetoacetic Acid AcetoneAcetyl-L-carnitine Alpha-Glucose Alpha-ketoisocaproic acidBenzoic AcidBetaine Beta-LactoseCitric Acid Creatine Creatinine D(-)FructoseD-(+)-Glyceric Acid D(+)-Xyl

10、oseDimethylamineDL-B-Aminoisobutyric Acid,DL-Carnitine DL-CitrullineDL-Malic AcidEthanolFormic AcidFumaric AcidGamma-Amino-N-Butyric AcidGamma-Hydroxybutyric Acid Gentisic AcidGlucoseGlutaric acid

11、GlycerolGlycineGlycolic AcidHippuric acidHomovanillic acidHypoxanthineImidazoleInositolisovaleric acidL(-) FucoseL-alanineL-asparagine L-aspartic acidL-Histidine,L-homocitrulline L-Isoleucine

12、L-Lactic AcidL-Lysine L-MethionineL-phenylalanineL-SerineL-ThreonineL-ValineMalonic AcidMethylamine Mono-methylmalonate N,N-dimethylglycine N-Butyric Acid Pimelic AcidPropionic AcidPyruvic Ac

13、id Salicylic acidSarcosineSuccinic AcidSucroseTaurineTrimethylamine Trimethylamine-N-Oxide Urea,血樣的NMR檢測(cè),,600MHz 1H-NMR spectra of rat plasma(A) NOESYPR1D, (B) CPMG, and (C) diffusion-edited spectrum,病人血

14、清的NMR檢測(cè),LC：肝癌患者；LD：一般肝病患者,,病人血清的NMR檢測(cè),,C：健康對(duì)照組；PC：胰腺癌患者,,,奶牛血清的NMR檢測(cè),,,A：臨床酮病奶牛；B：隱性酮病奶牛；C：正常對(duì)照組,牦牛血清的NMR檢測(cè),,,豬血清的NMR檢測(cè),,,尿樣的NMR檢測(cè),,Representative 600 MHz 1H-NMR spectra of urine from rat in (A) control and (B) dosed gro

15、up.,糞樣提取物的NMR檢測(cè),600 MHz 1H NMR spectra of rat fecal samples obtained from different groups,回腸沖洗物的NMR檢測(cè),,,600 MHz 1H NMR spectra of rat ileal flushes obtained from (A) control, (B) low-, (C) medium- and (D) high-dosage

16、groups,雞腸道內(nèi)容物的NMR檢測(cè),,,A：正常對(duì)照組；B：抗陰性菌組；C：抗陽(yáng)性菌組；D：廣譜抗菌組,羊水的NMR檢測(cè),,600 MHz 1H NMR spectra of rat amniotic fluids obtained from (A) control, (B) low-, (C) medium- and (D) high-dosage groups,腦積液的NMR檢測(cè),,組織樣的NMR檢測(cè),,Typical 600

17、 MHz HR-MAS CPMG 1H-NMR spectra of (A) kidney, (B) liver and (C) spleen from rat,豬腦的NMR檢測(cè),,,500 MHz 1H-HRMAS CPMG NMR spectra of pig brain,甲狀腺的NMR檢測(cè),,,組織提取物的NMR檢測(cè),,生物樣本的2D 核磁共振譜歸屬,Use of HSQC spectroscopy for analysis o

18、f common metabolites. In 1D spectra, overlapped signals hamper identification of individual metabolites, whereas in 2D correlation, spots are easily visible. (a) 1D 1H NMR spectrum of an equimolar mixture of the 26 sta

19、ndards. (b) 2D 1H–13C HSQC NMR spectra of the same synthetic mixture (red) overlaid onto a spectrum of aqueous whole-plant extract from Arabidopsis (blue).,HSQC used to select for protons directly bonded to 13C.,核磁共振譜圖

20、預(yù)處理,Bruker: TopspinAgilent: Vnmr-JMestReNova,核磁共振譜圖預(yù)處理,傅立葉變換和線寬因子（LB）,NMR信號(hào)一般都集中在FID的前部，后部含有的大部分為噪音。將FID乘一權(quán)重指數(shù)函數(shù)就能迫使尾部的FID為零，其程度由參數(shù)LB控制。使用這些函數(shù)雖可提高信噪比卻以犧牲分辨率為代價(jià)。,體液樣本推薦：1 Hz；組織樣本推薦：0.3 Hz,相位調(diào)整(Phasing),通常所采集到的譜圖含有吸收(abs

21、orption)與擴(kuò)散(dispersion)組份，通過(guò)相位調(diào)整可以的到純粹的吸收峰。,相位調(diào)整首先對(duì)最大峰進(jìn)行零級(jí)相位調(diào)整PH0，然后以一級(jí)相位調(diào)整PH1來(lái)調(diào)節(jié)其他的峰。,相位調(diào)整(Phasing),基線校正(baseline correct),多項(xiàng)式函數(shù) F(X)=A+B*X+C*X2+D*X3+E*X4Sine函數(shù) F(X)=A+B*sin(C*X+D)指數(shù)函數(shù) F(X)=A+B*exp(C*X),定標(biāo)

22、(Chemical shift reference),尿樣 TSP(d4) or DSS(d6): 0 ppm (-CH3) 單峰血樣 乳酸: 1.33 ppm(-CH3) 雙峰 α-葡萄糖: 5.22 ppm(CH1) 雙峰組織樣 乳酸: 1.33 ppm(-CH3) 雙峰 α-葡萄糖: 5.22 ppm(CH1) 雙峰,（A）尿樣檸檬酸區(qū)域的原始譜圖（B）譜峰對(duì)齊前的載荷圖；（C）譜峰對(duì)齊后

23、的載荷圖,峰對(duì)齊(peak alignment),排除干擾峰(Exclusion of resonances),水峰尿素峰加入的內(nèi)標(biāo)（如TSP）乙醇抗凝劑（如EDTA及其金屬配合物、檸檬酸）藥物代謝產(chǎn)物……,分段積分(Integral),積分區(qū)間積分間距（2-3 Hz）,歸一化(Normalization),相對(duì)于全譜積分面積歸一化相對(duì)于某一特定峰的峰面積歸一化(如TSP或肌酸) (有條件的定量分析)直接以分段積分的面

24、積進(jìn)行比較，不歸一化概率商歸一化,代謝組學(xué)中核磁共振數(shù)據(jù)的多變量統(tǒng)計(jì)分析,多變量統(tǒng)計(jì)分析,為什么要使用多變量統(tǒng)計(jì)分析,目的：采集生物樣本的譜圖，獲取各種生物信息,疾病 生理狀況 飲食 年齡 營(yíng)養(yǎng)狀況 基因變異 ……,人類(lèi)尿樣的NMR譜含有~10,000條譜峰 MS譜含有~百萬(wàn)條譜峰 需要多元統(tǒng)計(jì)投影的方法提取潛在的生物信息,LC-MS 或者NMR譜圖,為什么要使用多變量統(tǒng)計(jì)分析,多元統(tǒng)計(jì)投影方法有助于對(duì)高維復(fù)雜數(shù)據(jù),

25、對(duì)高維數(shù)據(jù)降維 減少?gòu)?fù)雜性 使數(shù)據(jù)可視化 可對(duì)樣本進(jìn)行分類(lèi)及預(yù)測(cè) 易于解釋 分離信號(hào)和噪聲,核磁共振數(shù)據(jù)前處理,標(biāo)度換算(Scaling),中心化換算(Mean center scaling)：Ctr自動(dòng)規(guī)格化(Unit variance scaling)：UV 其中帕萊托換算(Pareto scaling)：Par

26、 其中,,,,,,非監(jiān)督的模式識(shí)別(Unsupervised pattern recognition),主成分分析(Principal components analysis, PCA)系統(tǒng)聚類(lèi)分析(Hierarchical clustering analysis, HCA)非線性影射(Non-linear mapping, NLM)最小生成樹(shù)(Minimum spanning tree, MST ),U

27、nsupervisedMultivariate analysis based on projection methodsMain tool used in chemometricsExtract and display the systematic variation in the dataEach Principle Component (PC) is a linear combination of the original

28、data parametersEach successive PC explains the maximum amount of variance possible, not accounted for by the previous PCsPCs Orthogonal to each otherConversion of original data leads to two matrices, known as scores a

29、nd loadingsThe scores(T) represent a low-dimensional plane that closely approximates X. Linear combinations of the original variables. Each point represents a single sample spectrum. A loading plot/scatter plot(P) show

30、s the influence (weight) of the individual X-variables in the model. Each point represents a different spectral intensity. The part of X that is not explained by the model forms the residuals(E) X = TPT = t1p1T + t2p2T

31、 + ... + E,主成分分析(PCA),主成分分析(PCA),主成分分析(PCA),主成分分析(PCA),整體樣本總的代謝趨勢(shì)和輪廓排除樣本異常點(diǎn)用于質(zhì)量控制分析個(gè)體差異代謝軌跡疾病監(jiān)控,主成分分析的作用,偏最小二乘法-判別分析(Partial least squares-discriminant analysis, PLS-DA)貝葉斯概率論方法(Bayesian probabilistic approaches)線

32、性差別法(Linear discriminant analysis, LDA)神經(jīng)網(wǎng)絡(luò)(Neural networks, NN)簇類(lèi)獨(dú)立軟模式(Soft independent modeling of class analogy, SIMCA)偏最小二乘法 (Partial least squares, PLS),有監(jiān)督的模式識(shí)別(Supervised pattern recognition),PLS-DA,Supervised

33、 learning method. Principles that of PCA. But in PLS, a second piece of information is used, namely, the labeled set of class identities.Two data tables considered namely X (input data from samples) and Y (containing

34、qualitative values, such as class belonging, treatment of samples)?The quantitive relationship between the two tables is sought.X = TPT + EY = TCT + EThe PLS algorithm maximizes the covariance between the X variables

35、 and the Y variablesPLS models negatively affected by systematic variation in the X matrix not related to the Y matrix (not part of the joint correlation structure between X-Y.,偏最小二乘法-判別分析(Partial least squares-discrimi

36、nant analysis, PLS-DA),PLS-DA得分圖及以UV換算的載荷圖,OPLS-DA,OPLS method is a recent modification of the PLS method to help overcome pitfalls Main idea to seperate systematic variation in X into two parts, one linearly related t

37、o Y and one unrelated (orthogonal). Comprises two modeled variations, the Y-predictive (TpPpT) and the Y-orthogonal (ToPoT) compononents.Only Y-predictive variation used for modeling of Y. X = TpPpT + ToPoT + EY =

38、TpCpT + FE and F are the residual matrices of X and YOPLS-DA compared to PLS-DA,正交偏最小二乘法-判別分析(orthogonal Partial least squares-discriminant analysis, OPLS-DA),OPLS-DA得分圖及以UV換算并經(jīng)過(guò)回溯轉(zhuǎn)換后的載荷圖,OPLS-DA,組間的區(qū)分尋找潛在的生物標(biāo)志物比較分析

39、對(duì)未知樣本進(jìn)行預(yù)測(cè),(O)PLS-DA的作用,OPLS-DA中選擇生物標(biāo)志物的方式—VIP值法,Treated,Control,VIP臨界值的確定,取所有VIP值最高的1%，10%或20%取VIP值大于1,,,OPLS-DA中選擇生物標(biāo)志物的方式—相關(guān)系數(shù)法,相關(guān)系數(shù)臨界值的確定,相關(guān)系數(shù)臨界值表（單變量統(tǒng)計(jì)分析）由自己的樣本數(shù)及分析結(jié)果確定合適的臨界值 一般情況下，相關(guān)系數(shù)R>0.9為高度相

40、關(guān)，0.75<R<9為顯著相關(guān)，0.5<R<0.75為一般相關(guān),模型的交叉驗(yàn)證(Cross Validation),內(nèi)部驗(yàn)證(Internal validation)Q2: Q2>0.9: 預(yù)測(cè)能力很強(qiáng)； Q2>0.5: 預(yù)測(cè)能力較強(qiáng)； Q2更小時(shí)，預(yù)測(cè)能力一般。R2: R2 較高是得到較高 Q2的先決條件 最好同時(shí)具備較高的R2 和 Q2，且R2和Q2

41、的差別比較小。,模型的交叉驗(yàn)證(Cross Validation),外部驗(yàn)證(External model validation)舍一法(Leave one out method) 訓(xùn)練集，驗(yàn)證集，測(cè)試集排列實(shí)驗(yàn)(Permutation test): PLS-DACV-ANOVA: OPLS-DA,排列實(shí)驗(yàn)(Permutation test),(A)經(jīng)排列實(shí)驗(yàn)驗(yàn)證無(wú)意義的數(shù)據(jù)模型；(B)經(jīng)排列實(shí)驗(yàn)驗(yàn)證有意義的數(shù)據(jù)模型。,