Detection of Broad Bean Diseases by Fourier Transform Infrared Spectroscopy Combined with Curve Fitting

Xiaohua WANG，Gang LIU*，Quanhong OU，Xiangping ZHOU，Jianming HAO，Jianhong LIU，Luxiang WANG

1.School of Physics and Electronic Information,Yunnan Normal University,Kunming 650500,China;

2.Key Laboratory of Advanced Technique&Preparation for Renewable Energy Materials,Ministry of Education,Yunnan Normal University,Kunming 650500,China;

3.Institute of Quality Standards and Testing Technology,Yunnan Academy of Agricultural Sciences,Kunming 650223,China

Detection of Broad Bean Diseases by Fourier Transform Infrared Spectroscopy Combined with Curve Fitting

Xiaohua WANG1，Gang LIU1*，Quanhong OU1，Xiangping ZHOU1，Jianming HAO1，Jianhong LIU2，Luxiang WANG3

1.School of Physics and Electronic Information,Yunnan Normal University,Kunming 650500,China;

2.Key Laboratory of Advanced Technique&Preparation for Renewable Energy Materials,Ministry of Education,Yunnan Normal University,Kunming 650500,China;

3.Institute of Quality Standards and Testing Technology,Yunnan Academy of Agricultural Sciences,Kunming 650223,China

Fourier transform infrared(FTIR)spectroscopy was used to study diseased leaves in broad bean.Results showed that the infrared spectra of different broad bean diseased leaves were similar,which were mainly made up of the vibrational absorption bands of protein,lipid and polysaccharide.There were minor differences including the spectral peak position,peak shape and the absorption intensity in the range of 1 800-1 300 cm-1.There were obvious differences among their second derivative spectra in the range of 1 800-1 300 cm-1.After the procedure of the Fourier self-deconvolution and curve fitting of health bean leaves and broad bean diseased leaves in the range of 1 700-1 500 cm-1,three sub-peaks were obtained at 1 550 cm-1(protein amideⅡband),1 605 cm-1(lignin)and 1 650 cm-1(protein amide I band).The ratios of relative areas of the bands of amideⅡ,lignin,and amide I were 38.86%,28.68%and 32.47%in the spectra of healthy leaves,respectively.It was distinguished from the diseased leaves(chocolate spot leaf:15.42%, 42.98%and 41.61%,ring spot leaf:32.39%,35.63%and 31.98%,rust leaf:13.97%, 46.40%and 39.65%,yellowing leaf curl disease leaf:24.01%,36.55%and 39.44%). For sub-peak area ratios(A1563/A1605,A1650/A1605and A1563/A1654),those of four kinds of diseased leaves were smaller than that of healthy leaves,and there were also differences among four kinds of diseased leaves.The results proved that FTIR combining with curve fitting might be a potentially useful tool for detecting different kinds of broad bean diseases.

Fourier transform infrared(FTIR)spectroscopy;Broad bean diseases; Second derivative spectra;Curve fitting

B road bean is annual or biennial herb,can be used as grain, vegetable,fodder and green manure,and derives from Southwest Asia and North Africa.Broad bean contains eight kinds of essential amino acids,and carbohydrate content is 47%-60%.It has rich nutritive value. Broad bean contains abundant calcium,protein and phospholipid,and does not contain cholesterol,which is beneficial to human health.Tender broad bean has the effect neutralizing stomach and moistening intestinal tract;broad bean stem has the function of hematischesis and arresting diarrhea;broad bean flower has the effect of depressurization,arresting leucorrhoea and hematischesis[1].

Broad bean yield and quality are affected by plant diseases and insect pests,and plant disease is dominated by fungus disease and virus.Leaf diseases are mainly various leaf spot diseases,such as rust disease,ring spot disease,chocolate spot disease, brown spot disease,yellowing leaf curl disease(virus).The variety is various,

Common methods identifying crop disease include traditional biological method,microstructure observation and molecular biology method. These methods all require rich experience,skilled operation ability and long time.Fourier transform infrared(FTIR) spectroscopy is a kind of structure analysis technique based on vibrations of functional group and polar bond, can reflect vibrational mode of molecular functional group,has fingerprint characteristic,and has become forceful means detecting macromolecular structure and interaction in biological tissue.Infrared spectroscopy does not need separation and extraction process of sample,has the characteristics of simple operation,fast,low cost and non-pollution,and has been widely applied in studying plant and medicine[4-8]. Li Zhiyong et al.used FTIR to analyze the difference between diseased leaves(broad bean rust disease,stem base rot disease,ring spot disease and yellowing leaf curl disease)and healthy leaves,and used principal component analysis and clustering analysis to classify disease type of sample,which obtained better effect[9]. The purpose of this study is to discriminate different diseased leaves of broad bean by FTIR spectroscopy combined with curve fitting.

Materials and Methods

Materials

Broad bean leaf samples were collected from Luliang County,Qujing City,Yunnan Province during February-March of 2011 and 2012.After natural drying,sample was conserved to measure.When sampling,all samples avoided the vein,and we took the leaves at diseased spot.Frontier FTIR spectrometer from Perkin Elmer Company was used to measure spectrum.Spectral range was 4 000-400 cm-1;spectral resolution was 4 cm-1, and scanning times was 16.

Spectrum collection and data processing

Little sample and moderate KBr were put into agate mortar to grind. After grinding evenly,thin disc was pressed.Infrared spectrum was measured,and sample spectrum automatically deducted KBr background.Each sample collected five infrared absorption spectra,and the mean of five spectra was taken as infrared absorption spectrum of the sample.The spectrum was conducted by nine-point smoothing of OMNIC6.0 infrared spectrum software,base line correction and normalized preprocessing.In the range of 1 800-1 300 cm-1,software Origin 8.5 was used to obtain second-order derivative spectrum. Software Origin 8.5 was used to conduct curve fitting on original spectrum.

Results and Analysis

Infrared spectral characteristic analysis of broad bean leaves

Average infrared spectra of broad bean leaves were shown as Fig.1. Seen from Fig.1,major peaks of FTIR spectra for broad bean leaves included:there was a strong and wide absorption peak at 3 410 cm-1,which was mainly stretching vibration absorption of O-H[10];absorption peaks at 2 920 and 2 850 cm-1belonged to anti-symmetric and symmetric stretching vibrations of CH2[11];absorption peak at 2 954 cm-1belonged to anti-symmetric stretching vibration of-(CH3),which was mainly from tissue components of cell wall,such as protein,carbohydrate and lipid;absorption peak at 1 737 cm-1belonged to vibration absorption of lipid carboxyl group;absorption peak at 1 650 cm-1belonged to stretching vibration absorption of protein amide I band C=O,while absorption peak at 1 548 cm-1belonged to bending vibration of amide I band N=H and stretching vibration absorption of C-N;it was mixed vibration absorption region of protein,fat,lignin and polysaccharide between 1 500 and 1 200 cm-1[12],and peaks at 1 456 and 1 380 cm-1belonged to anti-symmetric and symmetric variable angle vibration absorption of methyl;absorption peak at 1 265 cm-1belonged to vibration absorption of protein amide III band[13];peak at 1 152 cm-1belonged to C-OH stretching vibration of polysaccharide, while peak near 1 075 cm-1belonged to C-O stretching vibration of polysaccharide.These spectral features reflected that major components of broad bean leaves were protein, polysaccharide and lipid.

Seen from Fig.1,infrared spectra of broad bean leaves infected by different diseases were roughly similar, and there was small difference in the range of 1 800-700 cm-1.Second derivative spectrum could amplify difference and improve spectral resolution.Fig.2 was second derivative spectra of different diseased leaves.In the 1 800-1 300 cm-1region,spectra of different diseased leaves were contrasted.It was clear that two strong peaks of leaves infected by broad bean rust disease(d)were respectively near 1 575 and 1 380 cm-1, and a strong peak of leaves infected by broad bean chocolate spot disease (b)was near 1 575 cm-1.Strong peak of healthy broad bean leaves(e)was near 1 470 cm-1,while other two kinds of leaves had a moderate-intensity absorption peak near 1 470 cm-1.In addition,peak intensity of broad bean leaves infected by rust disease(d)at 1 575 cm-1was stronger than that at 1 380 cm-1.Healthy broad bean leaves (e)had a strong peak near 1 540 cm-1, while other four kinds of leaves all presented very weak peaks at 1 530 cm-1. By comparing ring spot leaf(c)with yellowing leaf curl disease leaf(a),it was clear that absorption peak of broad bean leaves infected by ring spot disease(c)was stronger near 1 460 cm-1,while absorption peaks of broad bean leaves infected by yellowing leaf curl disease were stronger near 1 390 and 1 440 cm-1.It was clear that second derivative spectrum could distinguish broad bean leaves infected by different diseases.

Curve fitting

Intensity of infrared spectral absorption peak is often determined by peak height,peak width and peak intensity,and it is not enough accurate to use peak height to describe intensity.After Fourier self-deconvolution and curve fitting sub-peak processing,it is more close to actual situation to use peak area to represent its intensity[14-16]. Seen from Fig.2,the difference between healthy leaves and diseased leaves mainly concentrated between 1 700 and 1 500 cm-1.Curve fitting of infrared spectrum between 1 700-1 500 cm-1for different broad bean diseased leaves was conducted,and three sub-peaks were obtained(Fig.3). Absorption peak at 1 650 cm-1was mainly absorption of protein amide I band;absorption peak near 1 550 cm-1was mainly absorption of protein amide II band;it was absorption peak of lignin near 1 605 cm-1.Via area ratio of absorption peak(AamideI/Alignin, AamideII/Aligninand AamideII/AamideI),relative content change of protein in the leaves could be judged.

Table 1Sub-peak area ratios of diseased leaves and healthy leaves for broad bean

After Fourier self-deconvolution and curve fitting peak processing,yellowing leaf curl disease leaves obtained three sub-peaks at 1 559,1 610 and 1 654 cm-1.It was defined that anysub-peak area divided by area sum of three sub-peaks was area proportion of the sub-peak.Corresponding subpeak area proportions of yellowing leaf curl disease leaf were respectively 24.01%,36.55%and 39.44%;chocolate spot disease leaf obtained three sub-peaks at 1 547,1 611 and 1 656 cm-1,and corresponding sub-peak area proportions were respectively 15.42%,42.98%and 41.61%;ring spot leaf disease leaf obtained three sub-peaks at 1 565,1 611 and 1 654 cm-1,and corresponding sub-peak area proportions were respectively 32.39%,35.63%and 31.98%.Rust leaf had three sub-peaks at 1 546, 1 606 and 1 654 cm-1,and corresponding sub-peak area proportions were respectively 13.94%,46.60% and 39.65%;while healthy leaves had three sub-peaks at 1 563,1 605 and 1 654 cm-1,and corresponding sub-peak area proportions were respectively 38.86%,28.68%and 32.47%. At 1 605 cm-1,peak position,type and absorption intensity of each spectrum were all similar.Via area ratios of absorption peak(AamideI/Alignin,AamideII/Aligninand AamideII/AamideI),relative content change of protein in the diseased leaves could be judged.

Table 1 was peak area ratios of protein amide I band,amide II band and lignin in different broad bean diseased leaves.Seen from Table 1, AamideI/Aligninin healthy leaves was more than 1.1,and others were all less than 1.1,in which AamideI/Aligninof yellowing leaf curl disease leaf was close to 1.1; AamideII/Aligninof yellowing leaf curl disease leaf,chocolate spot disease leaf, ring spot disease leaf and rust disease leaf was all less than 1,while AamideII/Aligninof healthy leaf was more than 1.1;the maximum AamideII/AamideIwas 1.20 in healthy leaf,while the minimum was 0.35 of rust disease leaf.It showed that compared with healthy leaves,protein content in broad bean leaves infected by different diseases relatively declined,which indicated that area ratio difference of sub-peak could be used to distinguish different broad bean diseased leaves.

Conclusions

We studied FTIR spectra of four kinds of broad bean diseased leaves and healthy leaves.Results showed that major components of broad bean leaves were protein,lipid and polysaccharide.Their infrared spectra were overall similar at peak shape,intensity and position of footprint characteristic region.In the range of 1 700-1 500 cm-1,after Fourier self-deconvolution and curve fitting peak processing,area ratios of three sub-peaks(A1650/A1605, A1563/A1605and A1563/A1654)had obvious difference.Research showed that FTIR spectroscopy combining with curve fitting is a convenient,fast and nondestructive method for identifying different broad bean diseased leaves.

[1]ZHENG KB(鄭開(kāi)斌),LI AP(李?lèi)?ài)萍), ZANG CR(臧春榮),et al.Green legume vegetables——Broad bean(綠色豆類(lèi)蔬菜——蠶豆)[J].Xiandai Horticulture(現(xiàn)代園藝),2006(1):38-39.

[2]WANG XM(王曉鳴),ZHU ZD(朱振東), DUAN CX(段燦星),et al.Identification techniques of plant diseases and insect pests for broad bean and pea(蠶豆豌豆病蟲(chóng)害鑒別技術(shù))[M].Beijing:China Agricultural Science and Technology Press(北京:中國(guó)農(nóng)業(yè)科學(xué)技術(shù)出版社), 2007.

[3]OUYANG DS(歐陽(yáng)迪莎).Plant disease management in sustainable agriculture (可持續(xù)農(nóng)業(yè)中的植物病害管理)[D]. Fuzhou:Fujian Agriculture and Forestry University(福州:福建農(nóng)林大學(xué)),2005.

[4]YAN G,SONG P,SUN SQ,et al.Differentiation and quality estimation of Cordyceps with infrared spectroscopy [J].Spectrochimica Acta Part A:Molecular and Biomolecular Spectroscopy, 2009,74:983-990.

[5]LIU R,WANG M,DUAN JA,et al.Antipyretic and antioxidant activities of the aqueous extract of Cornububali(Water Buffalo Horn)[J].The American Journal of Chinese Medicine,2010,38:293-306.

[6]SUN SQ(孫素琴),ZHOU Q(周群), CHEN JB(陳建波).Infrared spectral analysis and identification of traditional Chinese medicine(中藥紅外光譜分析與鑒定)[M].Beijing:Chemical Industry Press(北京:化學(xué)工業(yè)出版社),2010.

[7]MATA PD,DOMINGUZ VA,SENDRA JMB,et al.Olive oil assessment in edible oil blends by means of ATR-FTIR and chemometrics[J].Food Control, 2012,23:449-455.

[8]SUN SQ(孫素琴),ZHOU Q(周群),QIN Z(秦竹).Two-dimensional related infrared spectral identification atlas of traditional Chinese medicine(中藥二維相關(guān)紅外光譜鑒定圖集)[M].Beijing: Chemical Industry Press(北京:化學(xué)工業(yè)出版社),2003.

[9]LI ZY(李志永),LIU G(劉剛),LI L(李倫), et al.FTIR research of diseased broad bean leaves(蠶豆病害葉的FTIR研究) [J].Spectroscopy and Spectral Analysis (光譜學(xué)與光譜分析),2012,32(5): 1217-1220.

[10]SHI YM(時(shí)有明),LIU G(劉剛),LIU JH (劉劍虹),et al.Study on black fungus from five different production places by FTIR spectroscopic methodology(利用傅里葉變換紅外光譜法研究了五種不同產(chǎn)地的黑木耳)[J].Acta Optica Sinica(光學(xué)學(xué)報(bào)),2007,27(1):129-132.

[11]DUBRAVKA K,MAJA B,MARIN K. FT-IR spectroscopy of lipoproteins——A comparative study[J]. Spectrochimica Acta Part A:Molecular and Biomolecular Spectroscopy,2009, 73:701-706.

[12]WENG SP(翁詩(shī)甫).FTIR spectroscope (傅里葉變換紅外光譜儀)[M].Beijing: Chemical Industry Press(北京:化學(xué)工業(yè)出版社),2005.

[13]LU XN,MOLLY AHW,MARIAH J,et al.A study of biochemical parameters associated with ovarian atresia and quality of caviar in farmed white sturgeon(Acipenser transmontanus)by Fourier Transform Infrared(FT-IR) Spectroscopy[J].Aquaculture,2011, 315:298-305.

[14]BYLER DM,SUSI H.Examination of the secondary structure of proteins by deconvolved FTIR spectra[J].Biopolymers,1986,25:469-487.

[15]KORKMAZ F,K?STER S,YILDIZ?, et al.In situ opening/closing of OmpG from E.coli and the splitting of β-sheet signals in ATR-FTIR spectroscopy[J]. Spectrochimica Acta Part A,2012,91 (6):395-401.

[16]SAMYN P,NIEUWKERKE VD,SCHOUKENS G,et al.Quality and statistical classification of brazilian vegetable oils using mid-infrared and raman spectroscopy[J].Applied Spectros copy,2012,66(5):552.

Responsible editor:Ping SONG

Responsible proofreader:Xiaoyan WU

基于曲線擬合的蠶豆病害葉的FTIR研究

汪小華1，劉剛1*，歐全宏1，周湘萍1，郝建明1，劉劍虹2，汪祿祥3
（1．云南師范大學(xué)物理與電子信息學(xué)院，云南昆明650500；2．云南師范大學(xué)可再生能源材料先進(jìn)技術(shù)與制備教育部重點(diǎn)實(shí)驗(yàn)室，云南昆明650500；3．云南省農(nóng)業(yè)科學(xué)院質(zhì)量標(biāo)準(zhǔn)與檢測(cè)技術(shù)研究所，昆明650223）

用傅里葉變換紅外光譜法研究蠶豆病害葉片，結(jié)果顯示不同病害蠶豆葉片紅外光譜圖整體相似，它們的紅外光譜主要由蛋白質(zhì)、脂類(lèi)和多糖的振動(dòng)吸收帶組成，僅在1 800～1 300 cm－1范圍光譜的峰位、峰形及吸收強(qiáng)度有一些微小差異。對(duì)1 800～1 300 cm－1波數(shù)范圍的光譜圖進(jìn)行二階導(dǎo)數(shù)處理，結(jié)果顯示蠶豆病害葉的二階導(dǎo)數(shù)譜差異明顯。對(duì)健康和病害蠶豆葉1 700～1 500 cm－1范圍光譜進(jìn)行傅里葉自去卷積和曲線擬合處理后，得到蛋白質(zhì)酰胺Ⅱ帶（1 550 cm－1）、木質(zhì)素（1 605 cm－1）和酰胺Ⅰ（1 650 cm－1）3個(gè)子峰，相應(yīng)子峰的峰面積比例顯示差異，黃化卷葉病分別為24．01%、36．55%、39．44%，赤斑病分別為15．42%、42．98%、41．61%，輪紋病分別為32．39%、35．63%、31．98%，銹病分別為13．97%、46．40%、39．65%，健康葉片分別為38．86%、28．68%、32．47%，健康葉的酰胺Ⅱ帶子峰相對(duì)面積比病害葉的大，而其木質(zhì)素子峰相對(duì)面積比病害葉的小。對(duì)于子峰面積比A1563／A1605、A1650／A1605和A1563／A1654，4種病害葉的比值均比健康葉的相應(yīng)數(shù)值小，4種病害葉之間也有差異。結(jié)果表明傅里葉變換紅外光譜（FTⅠR）結(jié)合曲線擬合可望對(duì)不同病害的樣品進(jìn)行有效鑒別。

傅里葉變換紅外光譜；蠶豆病害；二階導(dǎo)數(shù)光譜；曲線擬合and harm is great.They mainly harm leaf,sometimes stem,leaf handle and pod.If disease occurs in large area,serious economic loss can be caused[2].At present,plant disease prevention and control are dominated by chemical pesticide.Unreasonable utilization of pesticide and chemical fertilizer,drug resistance of harmful organisms and pesticide residue,etc. cause that eco-environment is continuously destroyed,and agricultural product quality declines.So,we should correctly identify disease and take reasonable prevention and control measures for corresponding disease,to reach sustainable development of crop eco-environment[3].

國(guó)家自然科學(xué)基金項(xiàng)目（30960179）；云南省高校科技創(chuàng)新團(tuán)隊(duì)支持計(jì)劃。

汪小華（1988－），女，安徽宿松人，碩士研究生，研究方向：生物光譜學(xué)，E-mail：zhaoxxiang1987＠sina．com。*通訊作者，E-mail：gliu66＠163．com。

2015-03-01

修回日期 2015-05-09

Supported by National Natural Science Foundation of China(30960179);Program for Innovative Research Team in Science and Technology in University of Yunnan Province.

*Corresponding author.E-mail:gliu66@163.com

Received:March 1,2015 Accepted:May 9,2015