CN107703173A - A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse - Google Patents
A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse Download PDFInfo
- Publication number
- CN107703173A CN107703173A CN201710483442.XA CN201710483442A CN107703173A CN 107703173 A CN107703173 A CN 107703173A CN 201710483442 A CN201710483442 A CN 201710483442A CN 107703173 A CN107703173 A CN 107703173A
- Authority
- CN
- China
- Prior art keywords
- sample
- standard
- nmr
- timber
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N24/00—Investigating or analyzing materials by the use of nuclear magnetic resonance, electron paramagnetic resonance or other spin effects
- G01N24/08—Investigating or analyzing materials by the use of nuclear magnetic resonance, electron paramagnetic resonance or other spin effects by using nuclear magnetic resonance
Landscapes
- Physics & Mathematics (AREA)
- High Energy & Nuclear Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
A kind of authentication method of the rare species timber based on 1H NMR datas storehouse, including the nuclear magnetic resonance nmr device sequentially connected, the N number of standard of wood sample of multiple seeds and testing sample of collection are measured;Phasing, baseline correction are carried out using MestReNova softwares, deducts solvent peak, subsection integral, normalized, are exported data, are established the 1H NMR datas storehouse of N × M structure of standard sample;Testing sample is analyzed to the classification and discriminating realized to testing sample compared with the data of standard sample.The present invention to multiple standard of wood samples of collection by carrying out 1H NMR spectra measure, analyzed using sPLS DA, success is made a distinction wood sample by kind, realizes the classification to unknown sample, being finally introducing computer system can automatically process to obtain visualization result.Different types of timber is distinguished well on analysis result figure, a kind of new method is provided for the taxonomic history of timber.
Description
Technical field
It is especially a kind of to use computer auxiliaring means, based on 1H-NMR numbers the present invention relates to a kind of Wood Identification Method
According to the identification apparatus and method in storehouse.
Background technology
The earlier processes of timber identification mainly rely on artificial experience and knowledge, and combine some aids, such as retrieve
Table, verge-perforated card, and for example the day for announcing is on October 1st, 1986, and notification number is CN86200457U Chinese patent application
Disclosed number timber identification card, etc., according to gross feature and microscopic feature, progressively identified by observing, comparing and analyze
Identify timber, this method is influenceed by the subjective factor of assessor, and very big, discrimination is low, cause some precious, prices compared with
Timber high or with specific use, is often taken as and is typically handled with material, as typically timber is used, also occurs because knowing
It is not unclear and cause the blindly phenomenon that uses, misuse, not only to wood producing, sale and using etc. links make troubles,
The raising of forest industrial enterprise economic interests is have impact on, wastes the resource of preciousness, further, since being consumed by artificial recognition methods
It is time-consuming very long, the waste of human resources is also caused to a certain extent.
In recent years, area of computer aided timber identification is increasingly becoming a kind of Wood Identification Method of main flow, such as authorized announcement date
For on August 5th, 2009, the near infrared light spectrum discrimination side for the redwood that notification number is announced by 100523793C Chinese patent
Method, it is exactly using redwood and non-redwood wood sample, is adopted using diverse location of the near infrared spectrum equipment on wood sample surface
Collection near infrared spectrum for several times, spectrum is gathered respectively 3~10 positions for same sample, then by Pretreated spectra,
After smooth, baseline correction, first derivative, second dervative, multiplicative scatter correction or Data Dimensionality Reduction pretreatment, independently built by soft
The multivariate data analysis method such as mould classification or least square discriminant analysis, establish the differentiation mould of true and false redwood and its timber varieties of trees
Type, so as to which the model using foundation is the differentiation that can be achieved why to add redwood and timber varieties of trees.This method has higher analysis
Efficiency and the degree of accuracy, but due to its based on sample be timber material object, the timber for being not easy to sample for some and carrying, in reality
Operating aspect, there is certain difficulty.Based on lumber fibre or stereoscopic image, using the recognition methods of machine vision means
Overcome above mentioned problem, it is well known that timber identification is often determined by the key feature region in timber picture, and timber figure
In piece containing fracture, resin canal, the feature such as damage by worms, go mouldy noise region, then the identification to timber has certain suppression work
With existing such recognition methods can not well be eliminated and made an uproar without the effect for effectively protruding timber key feature region
The influence in sound area domain, so as to which discrimination greatly reduce.
NMR is widely used in structural analysis and performance study, the molecule knot in the fields such as pharmacy, chemistry, food, bioengineering
The repercussion study of structure, such as to organic compound, Chinese herbal medicine effective ingredientses, inorganic matter, polymer, high-molecular compound, gold
Belong to organic compound and complex compound etc. and carry out structure determination;To organ or tissue, micro metabolic components detect into the cell, are
The research of the natural macromoleculars such as polypeptide, sugar, enzyme, ribonucleic acid, cellulose and micro biological sample provides much information;Using
Line HPLC-NMR improves signal resolution and peak resolution, and the drug metabolite that sample is collected for epidemiology purpose is reflected
The influence of biomarker identification and research probiotics to simulating mammlian system fixed, in toxicity assessment etc..Moreover,
NMR is not only used alone, and is also combined with mass spectrum for the toxicity research in metabolism group.Dependency relation analysis method is tied with NMR
Close the proteomics data for being also extended to one model of human prostate carcinoma of analysis, and mouse angiocardiopathy and obese model
Genomics data.
As cell biology, molecular biology, science of heredity are developed rapidly and genetic marker research goed deep into, NMR
Condition is provided for research metabolism and Physiology and biochemistry.NMR plays more next in the research of bioinformatics and structure biology
More important effect.The continuous development of NMR technology and maturation, its application field is widened, such as from protein, polypeptide, nucleic acid
Structure of biological macromolecule and their interactions with small molecule are to process mechanisms such as the physiology of animal and human body, pathology, biochemistry
Research, it may also be used for drug toxicity evaluation, the diagnosis of clinical patients disease, the detection of therapeutic effect etc..Nuclear magnetic resonance exists
The advantages of in metabolism research is exactly the simplification of non-invasi and sample pre-treatments, but its weak point is then that muting sensitivity is low
Resolution ratio.
The content of the invention
It is an object of the invention to overcome in the prior art discrimination compared with, reliability is low the problem of, it is proposed that one kind has
The identification apparatus and method of higher discrimination and the rare species timber based on 1H-NMR databases of reliability.
To reach above-mentioned purpose, the technical solution adopted in the present invention is:
A kind of identification apparatus of the rare species timber based on 1H-NMR databases, including the nuclear magnetic resonance nmr dress sequentially connected
Put, ultrosonic extractor, the computer installation of band model identification software, the nuclear magnetic resonance(NMR)Device includes sequentially connecting
Connect
The magnet of one generation high-performance constant fixed-field, there is provided magnetic field uniformly, stable, to ensure the high-resolution of experimental method
And high sensitivity;The radio system of one generation exciting field, to control the transmitting of radio-frequency pulse;
One is responsible for the probe that signal is radiated in transmitting excitation pulse to sample and reception;
One enhancing and the detecting system of detection spin response, include preamplifier and receiver, the detecting system will altogether
The signal that shakes is converted into sensor current signal, and appropriate amplification is carried out to signal according to gain);
And computer and control system;
With some sample temperature control units, the servicing unit of air compressor machine;
Wherein, under the stationary magnetic field environment that superconducting magnet provides, radio system control probe transmitting full range radio-frequency pulse is different
The proton of precession frequency resonates under corresponding radio-frequency pulse frequency, and caused resonance signal gathers through probe, then via
Detecting system conversion, zoom into sensor current signal(FID)Reach spectrometer.Computer is by sensor current signal(FID)Signal turns
Change resonance spectrogram into.
A kind of authentication method of the rare species timber based on 1H-NMR databases, comprises the following steps:
1. pair N number of standard of wood sample of multiple seeds collected is measured, wherein using each standard sample of pure Methanol determination
Chromatogram, using the spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample, the proton nmr spectra to get standard samples;
2. phasing, baseline correction, button are carried out successively to above-mentioned proton nmr spectra spectrogram using MestReNova softwares
Except solvent peak, subsection integral, normalized, data are exported, obtain chemical shift and associated quad area;By above-mentioned signal
After processing, the 1H-NMR databases of N × M structure of standard sample are established;
3. a pair unknown timber testing sample is measured, wherein using the chromatogram of each unknown testing sample of pure Methanol determination,
The spectroscopic signal of unknown testing sample is determined using superconduction nuclear magnetic resonance spectrometer, obtains the nuclear magnetic resonance of unknown testing sample
Hydrogen is composed;Carry out phasing, baseline correction successively to spectrogram using MestReNova softwares, deduct solvent peak, subsection integral, return
One change is handled, and is exported data, is obtained chemical shift and associated quad area;Obtain unknown treating test sample by above-mentioned signal transacting
The 1H-NMR data of product;
4. the 1H-NMR data of testing sample are analyzed into realization to unknown sample compared with the data of 1H-NMR databases
Sort out and differentiate.
Preferably, following optimization method is included using the chromatogram of each standard sample of pure Methanol determination in above-mentioned steps 1:
11. the influence of the amount and the aspect factor of ultrasonic extraction time two of deuterated methanol to signal to noise ratio is added by investigating, to confirm
Optimal extracting method, determine deuterated methanol addition 1ml, ultrasonic extraction time 30min.
12. electricity consumption is drilled on wood sample uniformly drilling 4~6, powder timber sample is taken out, is mixed.Weigh 100mg in
In 2mL centrifuge tubes, 1ml deuterated methanols, ultrasonic extraction 30min, high speed centrifugation 10min are added(The rpm of rotating speed 13000), in transfer
Clear liquid determines proton nmr spectra, obtained containing more, the stronger information in peak more horn of plenty, stability more to 5mm bore nuclear magnetic tubes
By force, stronger standard sample collection of illustrative plates --- the proton nmr spectra of repeatability.
Preferably, included in above-mentioned steps 1 using the spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample following
Method and condition:Standard sample is measured using standard one-dimensional zgpr pulses, major parameter is set:Spectrum width 16ppm, sampling
Point 64k, sampling time 4s, scanning times 24, by the M standard of wood sample of collected multiple seeds, electricity consumption is drilled in timber
Upper uniformly drilling 4~6, takes out powder-like, mixes.100mg powder is weighed in 2mL centrifuge tubes, 1ml deuterated methanols is added, surpasses
Sound extracts 30min, high speed centrifugation 10min(Rotating speed 13000rpm), transfer supernatant is to 5mm bore nuclear magnetic tubes, to standard sample
Carry out the measure of proton nmr spectra.
Preferably, phase is carried out successively to above-mentioned proton nmr spectra spectrogram using MestReNova softwares in above-mentioned steps 2
Bit correction, baseline correction, solvent peak is deducted, subsection integral, normalized, export data, obtain chemical shift and corresponding product
Facet is accumulated, and it comprises the following steps:
21. the master sample information known to is classified to sample, by the master sample data of the standard timber of Given information
Collection is grouped according to different type;
22. be modeled according to the classification of step 21, according to the weight of variable and the correlation calculations variable of set classification,
Retain and contribute classification big variable, by penalty, make clothes
FromReduced model is calculated, ridge regression makes to contribute classification small factor beta j=0, i.e. the weight of variable is zero, from
And ignored and be not used in modeling;Comprise the following steps that:Variable is selectively put into model so as to obtain more preferable property by LASSO
Can parameter.According to the parameter set, allow algorithm to select several different variables automatically, fit the different model of coefficient.
It is worth noting that, algorithm only receives numerical matrix as mode input, it is necessary to will if having discrete variable in independent variable
This row discrete variable is converted into the vector that several row comprise only 0 and 1.Thus, penalty accounting equation be linear inequality about
The quadratic programming problem of beam.Adjustment parametert>=0 controls the quantity for being used for Shrinkage estimation.OrderEstimate for complete least square
Meter, order。t<t 0Solution will be caused to be compressed to 0 direction, and cause some coefficients to be equal to 0;
23. pair variable retained carries out partial least squares discriminant analysis, the non-zero weight variable of reservation is subjected to multivariable
PLS regression analyses:
By formulaPrincipal component and load vectors is calculated;
24. drawing the shot chart of PLS-DA analyses, sample can be carried out into two-dimensional projection according to discriminant analysis obtains sPLS-DA
Shot chart, identical group of sample is adjacent to each other, and difference group sample is located remotely from each other;
25. drawing fiducial confidence ellipse according to sample distribution situation, the help file according to mixOmics bags calculates confidential interval, writes from memory
Think 0.95, and the ellipse that confidence level is 95% is drawn on shot chart;
26. sample to be determined is added into standard data set, with master sample together modeling of class, unknown sample data are added
Enter in master sample data set;
27. judge whether similar with one of which sample
(a)Fall into the fiducial confidence ellipse of one of which sample
(b)Without falling into the fiducial confidence ellipse of any group of sample.
Brief description of the drawings
Fig. 1 is the basic structure composition figure of H NMR spectroscopy instrument of the present invention;
Fig. 2 and Fig. 3 is that unknown sample sandalwood X and two groups of standard samples of any combination are carried out into analysis result schematic diagram,
A kind of identification apparatus of rare species timber based on 1H-NMR databases of the present invention, including the core sequentially connected
Magnetic resonance NMR devices, ultrosonic extractor, the computer installation of band model identification software, the nuclear magnetic resonance(NMR)Device
Including what is sequentially connected:
The magnet of one generation high-performance constant fixed-field, there is provided magnetic field uniformly, stable, to ensure the high-resolution of experimental method
And high sensitivity;The radio system of one generation exciting field, to control the transmitting of radio-frequency pulse;
One is responsible for the probe that signal is radiated in transmitting excitation pulse to sample and reception;
One enhancing and the detecting system of detection spin response, include preamplifier and receiver, the detecting system will altogether
The signal that shakes is converted into sensor current signal, and appropriate amplification is carried out to signal according to gain;
And computer and control system;
With some sample temperature control units, the servicing unit of air compressor machine;
Wherein, under the stationary magnetic field environment that superconducting magnet provides, radio system control probe transmitting full range radio-frequency pulse is different
The proton of precession frequency resonates under corresponding radio-frequency pulse frequency, and caused resonance signal gathers through probe, then via
Detecting system conversion, zoom into sensor current signal(FID)Reach spectrometer.Computer is by sensor current signal(FID)Signal turns
Change resonance spectrogram into.
A kind of authentication method of rare species timber based on 1H-NMR databases of the present invention, comprises the following steps:
1. pair N number of standard of wood sample of multiple seeds collected is measured, wherein using each standard sample of pure Methanol determination
Chromatogram, using the spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample, the proton nmr spectra to get standard samples;
2. phasing, baseline correction, button are carried out successively to above-mentioned proton nmr spectra spectrogram using MestReNova softwares
Except solvent peak, subsection integral, normalized, data are exported, obtain chemical shift and associated quad area;By above-mentioned signal
After processing, the 1H-NMR databases of N × M structure of standard sample are established;
3. a pair unknown timber testing sample is measured, wherein using the chromatogram of each unknown testing sample of pure Methanol determination,
The spectroscopic signal of unknown testing sample is determined using superconduction nuclear magnetic resonance spectrometer, obtains the nuclear magnetic resonance of unknown testing sample
Hydrogen is composed;Carry out phasing, baseline correction successively to spectrogram using MestReNova softwares, deduct solvent peak, subsection integral, return
One change is handled, and is exported data, is obtained chemical shift and associated quad area;Obtain unknown treating test sample by above-mentioned signal transacting
The 1H-NMR data of product;
4. the 1H-NMR data of testing sample are analyzed into realization to unknown sample compared with the data of 1H-NMR databases
Sort out and differentiate.
Preferably, following optimization method is included using the chromatogram of each standard sample of pure Methanol determination in above-mentioned steps 1:
11. the influence of the amount and the aspect factor of ultrasonic extraction time two of deuterated methanol to signal to noise ratio is added by investigating, to confirm
Optimal extracting method, determine deuterated methanol addition 1ml, ultrasonic extraction time 30min.
12. electricity consumption is drilled on wood sample uniformly drilling 4~6, powder timber sample is taken out, is mixed.Weigh 100mg in
In 2mL centrifuge tubes, 1ml deuterated methanols, ultrasonic extraction 30min, high speed centrifugation 10min are added(The rpm of rotating speed 13000), in transfer
Clear liquid determines proton nmr spectra, obtained containing more, the stronger information in peak more horn of plenty, stability more to 5mm bore nuclear magnetic tubes
By force, stronger standard sample collection of illustrative plates --- the proton nmr spectra of repeatability.
Preferably, included in above-mentioned steps 1 using the spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample following
Method and condition:Standard sample is measured using standard one-dimensional zgpr pulses, major parameter is set:Spectrum width 16ppm, sampling
Point 64k, sampling time 4s, scanning times 24, by the M standard of wood sample of collected multiple seeds, electricity consumption is drilled in timber
Upper uniformly drilling 4~6, takes out powder-like, mixes.100mg powder is weighed in 2mL centrifuge tubes, 1ml deuterated methanols is added, surpasses
Sound extracts 30min, high speed centrifugation 10min(Rotating speed 13000rpm), transfer supernatant is to 5mm bore nuclear magnetic tubes, to standard sample
Carry out the measure of proton nmr spectra.
Preferably, phase is carried out successively to above-mentioned proton nmr spectra spectrogram using MestReNova softwares in above-mentioned steps 2
Bit correction, baseline correction, solvent peak is deducted, subsection integral, normalized, export data, obtain chemical shift and corresponding product
Facet is accumulated, and it comprises the following steps:
21. the master sample information known to is classified to sample, by the master sample data of the standard timber of Given information
Collection is grouped according to different type;
22. be modeled according to the classification of step 21, according to the weight of variable and the correlation calculations variable of set classification,
Retain and contribute classification big variable, by penalty, make clothes
FromReduced model is calculated, ridge regression makes to contribute classification small factor beta j=0, i.e. the weight of variable is zero, from
And ignored and be not used in modeling;Comprise the following steps that:LASSO is not that all variables are all put into model to be fitted,
But variable is selectively put into model so as to obtain better performance parameter.According to the parameter set, make algorithm automatic
Several different variables are selected, fit the different model of coefficient.It is worth noting that, algorithm can only receive numerical matrix work
For mode input, it is necessary to which this row discrete variable, which is converted into several row, comprises only 0 He if having discrete variable in independent variable
1 vector.Thus, penalty accounting equation is the quadratic programming problem of linear inequality constraint.Adjustment parametert>=0 control
The quantity for Shrinkage estimation.OrderFor complete least-squares estimation, order。t<t 0Solution party will be caused
Case is compressed to 0 direction, and causes some coefficients to be equal to 0,
23. pair variable retained carries out partial least squares discriminant analysis, the non-zero weight variable of reservation is subjected to multivariable
PLS regression analyses:
By formulaPrincipal component and load vectors is calculated;
24. drawing the shot chart of PLS-DA analyses, sample can be carried out into two-dimensional projection according to discriminant analysis obtains sPLS-DA
Shot chart, identical group of sample is adjacent to each other, and difference group sample is located remotely from each other;
25. drawing fiducial confidence ellipse according to sample distribution situation, the help file according to mixOmics bags calculates confidential interval, writes from memory
Think 0.95, and the ellipse that confidence level is 95% is drawn on shot chart;
26. sample to be determined is added into standard data set, with master sample together modeling of class, unknown sample data are added
Enter in master sample data set;
27. judge whether similar with one of which sample
(a)Fall into the fiducial confidence ellipse of one of which sample
(b)Without falling into the fiducial confidence ellipse of any group of sample.
Embodiment 1:
1st, testing sample is measured
1.1 embodiment materials:
1.1.1 the present embodiment collects 9 seeds, 73 standard of wood samples altogether(Wherein 8 most common, most confusing sandalwood trees
Kind, standard sample 67;Another shellac varnish wood is not sandalwood, but is commonly used for counterfeit foregoing sandalwood, therefore is tested herein in the lump,
Standard sample 6), identify and approve by quality surveillance forestry products inspection point of Guangdong Province.
1.1.2 Chromatographic Pure Methanol.Acetonitrile is bought in Tedia companies of the U.S., and deuterated methanol is bought to be implemented in Cambridge Isotope
Example room.
1.1.3 nuclear magnetic resonance hydrogen spectruming determining is determined using the 500MHz superconduction nuclear magnetic resonance spectrometers of Bruker AVANCE III.
1.2 methodologies are confirmed
1.2.1 the optimization of extracting method
The influence of the amount and the aspect factor of ultrasonic extraction time two of deuterated methanol to signal to noise ratio is added by investigating, it is optimal to confirm
Extracting method.Measurement result shows, wherein adding solvent peak water peak height optimum when deuterated methanol amount is 1mL;Ultrasonic extraction
30min is compared with ultrasonic 15min, and ultrasonic 30min signals are clearly more powerful, and the signal of subregion is more, and the ultrasonic extraction time
The sample collection of illustrative plates substantially indifference that 45min and 30min is obtained, show to have extracted at 30 minutes fully, reach saturation.Therefore determine
Final scheme is deuterated methanol addition 1ml, ultrasonic extraction time 30min.
1.2.2 the optimization of sampling method
Electricity consumption is drilled on sandalwood sample uniformly drilling 4~6, takes out powder sandalwood sample, mixes.100mg is weighed in 2mL centrifuge tubes
In, add 1ml deuterated methanols, ultrasonic extraction 30min, high speed centrifugation 10min(The rpm of rotating speed 13000), transfer supernatant to 5mm
Bore nuclear magnetic tube, determine proton nmr spectra.Measurement result shows that (contained peak is more for the sample collection of illustrative plates that the sampling method obtains
It is more, stronger) information more horn of plenty.
1.2.3 replica test
By carrying out 6 sub-sampling measure to same sample, obtained collection of illustrative plates is subjected to similarity analysis to investigate this method
Repeatability.Test result shows, carries out similarity analysis to the proton nmr spectra of 6 parts of samples, as a result reaches more than 0.99,
Show that the repeatability for changing method is preferable.
1.2.4 sample stability is tested
By once being determined every 12 hours to same sample, determine 6 times altogether, obtained collection of illustrative plates is subjected to similarity analysis
To investigate the stability of this method.Test result shows, carries out similarity analysis to obtained proton nmr spectra, as a result reaches
To more than 0.99, show that sample is preferable in 72 hours internal stabilities.
The condition of 1.3 nuclear magnetic resonance hydrogen spectruming determinings
Sample is measured using standard one-dimensional zgpr pulses.Major parameter is set:Spectrum width 16ppm, sampled point 64k, sampling
Time 4s, scanning times 24.
1.4 by 73 standard of wood samples of 9 collected seeds, and electricity consumption is drilled on timber uniformly drilling 4~6,
Powder-like is taken out, is mixed.100mg powder is weighed in 2mL centrifuge tubes, adds 1ml deuterated methanols, ultrasonic extraction 30min, at a high speed from
Heart 10min(Rotating speed 13000rpm), transfer supernatant is for testing sample, carries out nuclear-magnetism to sample and be total to 5mm bore nuclear magnetic tubes
The measure of the hydrogen that shakes spectrum.
2nd, sandalwood database is established
2.1 are carried out phasing, baseline correction using MestReNova softwares, deduct solvent peak successively to spectrogram, segmentation product
Point, normalized, data are exported, obtain chemical shift and associated quad area.
After above-mentioned signal transacting, the 1H-NMR databases of 73 × 186 structures as shown in the table are established.It is i.e. whole
Individual database includes 73 samples, and each sample includes 186 data points, altogether individual data more than 13500.
2.2 sPLS-DA method data analysis flowcharts
2.2.1 master sample data set is grouped according to different type
2.2.2 by penaltyReduced model is calculated, ridge regression makes to tribute of classifying
Small factor beta j=0 is offered, i.e. the weight of variable is zero, and modeling is not used in so as to be ignored
2.2.3 carries out the non-zero weight variable of reservation the PLS regression analyses of multivariable:
By formulaPrincipal component and load vectors is calculated
2.2.4 sample can be carried out by two-dimensional projection according to discriminant analysis and obtains sPLS-DA shot charts, identical group of sample leans on mutually
Closely, different group samples are located remotely from each other
2.2.5 confidential interval is calculated(It is defaulted as 0.95), and the ellipse that confidence level is 95% is drawn on shot chart
2.2.6 position sample data are added in master sample data set
2.2.7 it is whether similar with one of which sample
(a)Fall into the fiducial confidence ellipse of one of which sample
(b)Without falling into the fiducial confidence ellipse of any group of sample
3rd, the data of testing sample are analyzed compared with the data of database
3.1 pairs of 73 sandalwood samples carry out 1H-NMR collection of illustrative plates measure, carry out chemometrics application to the data obtained, as a result table
It is bright to use sparse partial least squares discriminant analysis(sPLS-DA), successfully sandalwood sample can be made a distinction by kind, and realization pair
The classification of unknown sample, it is with a high credibility up to more than 95%.
3.2 embodiment results show that different types of sandalwood is distinguished well on analysis result figure, illustrate institute
The method of foundation has powerful classification and discriminating effect, provides a kind of new method for the taxonomic history of sandalwood, and can be
Establish brand-new timber discrimination method and provide model and basis.
3.3 Fig. 2 and Fig. 3 are that unknown sample sandalwood X and two groups of standard samples of any combination are carried out into analysis result.Such as figure
Shown in 2, unknown sample and Ovshinsky yellow wingceltis, Bali yellow wingceltis group sPLS-DA analysis of two-dimensional, graphics, unknown sample fall in each group 95%
Outside confidential interval, the results showed that be not belonging to this two groups.As shown in figure 3, unknown sample and Ovshinsky yellow wingceltis, hedgehog red sandalwood group sPLS-
DA analysis of two-dimensional, graphics, unknown sample fall outside the confidential interval of each group 95%, the results showed that are not belonging to this two groups.
Claims (5)
1. a kind of identification apparatus of the rare species timber based on 1H-NMR databases, it is characterised in that including the core sequentially connected
Magnetic resonance NMR devices, ultrosonic extractor, the computer installation of band model identification software, the nuclear magnetic resonance(NMR)Device
Including what is sequentially connected:
The magnet of one generation high-performance constant fixed-field, there is provided magnetic field uniformly, stable, to ensure the high-resolution of experimental method
And high sensitivity;The radio system of one generation exciting field, to control the transmitting of radio-frequency pulse;
One is responsible for the probe that signal is radiated in transmitting excitation pulse to sample and reception;
One enhancing and the detecting system of detection spin response, include preamplifier and receiver, the detecting system will altogether
The signal that shakes is converted into sensor current signal, and appropriate amplification is carried out to signal according to gain;
And computer and control system;
With some sample temperature control units, the servicing unit of air compressor machine;
Wherein, under the stationary magnetic field environment that superconducting magnet provides, radio system control probe transmitting full range radio-frequency pulse is different
The proton of precession frequency resonates under corresponding radio-frequency pulse frequency, and caused resonance signal gathers through probe, then via
Detecting system conversion, zoom into sensor current signal(FID)Reach spectrometer, computer is by sensor current signal(FID)Signal turns
Change resonance spectrogram into.
2. a kind of authentication method of the rare species timber based on 1H-NMR databases, it is characterised in that comprise the following steps:
The N number of standard of wood sample of multiple seeds of collection is measured, wherein the color using each standard sample of pure Methanol determination
Spectrum, using the spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample, the proton nmr spectra to get standard samples;
Above-mentioned proton nmr spectra spectrogram is carried out successively using MestReNova softwares phasing, baseline correction, deduct it is molten
Agent peak, subsection integral, normalized, data are exported, obtain chemical shift and associated quad area;By above-mentioned signal transacting
Afterwards, the 1H-NMR databases of N × M structure of standard sample are established;
Unknown timber testing sample is measured, wherein using the chromatogram of each unknown testing sample of pure Methanol determination, adopted
The spectroscopic signal of unknown testing sample is determined with superconduction nuclear magnetic resonance spectrometer, obtains the hydrogen nuclear magnetic resonance of unknown testing sample
Spectrum;Carry out phasing, baseline correction successively to spectrogram using MestReNova softwares, deduct solvent peak, subsection integral, normalizing
Change is handled, and is exported data, is obtained chemical shift and associated quad area;Unknown testing sample is obtained by above-mentioned signal transacting
1H-NMR data;
The 1H-NMR data of testing sample are analyzed into realization compared with the data of 1H-NMR databases to return unknown sample
Class and discriminating.
3. the authentication method of the rare species timber according to claim 2 based on 1H-NMR databases, it is characterised in that
The chromatogram of each standard sample of pure Methanol determination is being used to include following optimization method in above-mentioned steps 1:
The influence of the amount and the aspect factor of ultrasonic extraction time two of deuterated methanol to signal to noise ratio is added by investigating, it is optimal to confirm
Extracting method, deuterated methanol addition 1ml, ultrasonic extraction time 30min are determined,
12. electricity consumption be drilled on wood sample uniformly drilling 4~6, take out powder timber sample, mix, weigh 100mg in 2mL from
In heart pipe, 1ml deuterated methanols, ultrasonic extraction 30min, high speed centrifugation 10min are added(The rpm of rotating speed 13000), shift supernatant
To 5mm bore nuclear magnetic tubes, determine proton nmr spectra, obtain containing more, the stronger information in peak more horn of plenty, stability is stronger,
Stronger standard sample collection of illustrative plates --- the proton nmr spectra of repeatability.
4. the authentication method of the rare species timber according to claim 2 based on 1H-NMR databases, it is characterised in that
The spectroscopic signal of superconduction nuclear magnetic resonance spectrometer bioassay standard sample is being used to include following methods and condition in above-mentioned steps 1:
Standard sample is measured using standard one-dimensional zgpr pulses, major parameter is set:Spectrum width 16ppm, sampled point 64k, sampling
Time 4s, scanning times 24, by the M standard of wood sample of collected multiple seeds, electricity consumption is drilled on timber uniformly drilling 4
~6, powder-like is taken out, is mixed, weighs 100mg powder in 2mL centrifuge tubes, adds 1ml deuterated methanols, ultrasonic extraction 30min,
High speed centrifugation 10min(Rotating speed 13000rpm), transfer supernatant to 5mm bore nuclear magnetic tubes, nuclear magnetic resonance is carried out to standard sample
The measure of hydrogen spectrum.
5. the authentication method of the rare species timber according to claim 2 based on 1H-NMR databases, it is characterised in that
Using MestReNova softwares above-mentioned proton nmr spectra spectrogram is carried out successively in above-mentioned steps 2 phasing, baseline correction,
Solvent peak is deducted, subsection integral, normalized, data is exported, obtains chemical shift and associated quad area, it includes following
Step:
Sample is classified according to known master sample information, the master sample data set of the standard timber of Given information is pressed
It is grouped according to different type;
It is modeled according to the classification of step 21, according to the weight of variable and the correlation calculations variable of set classification, is retained
Big variable is contributed classification, by penalty, make obedienceReduced model is calculated, ridge regression makes to contribute classification small factor beta j=0, i.e. the weight of variable is zero, so as to
Ignored and be not used in modeling;Comprise the following steps that:Variable is selectively put into model so as to obtain better performance by LASSO
Parameter, according to the parameter set, allow algorithm to select several different variables automatically, fit the different model of coefficient, this
When, algorithm only receives numerical matrix as mode input, if having discrete variable in independent variable, it is necessary to this row are discrete
Variables transformations comprise only 0 and 1 vector for several row, thus, penalty accounting equation is the secondary rule of linear inequality constraint
The problem of drawing, adjustment parametert>=0 controls the quantity for being used for Shrinkage estimation, orderFor complete least-squares estimation, order,t<t 0Solution will be caused to be compressed to 0 direction, and cause some coefficients to be equal to 0;
Partial least squares discriminant analysis is carried out to the variable of reservation, the PLS that the non-zero weight variable of reservation is carried out to multivariable is returned
Return analysis:
By formulaPrincipal component and load vectors is calculated;
The shot chart of PLS-DA analyses is drawn, sample can be carried out into two-dimensional projection according to discriminant analysis obtains sPLS-DA scores
Figure, identical group of sample is adjacent to each other, and difference group sample is located remotely from each other;
Fiducial confidence ellipse is drawn according to sample distribution situation, the help file according to mixOmics bags calculates confidential interval, is defaulted as
0.95, and the ellipse that confidence level is 95% is drawn on shot chart;
Sample to be determined is added into standard data set, with master sample together modeling of class, unknown sample data is added and marked
Quasi- sample data is concentrated;
Judge whether similar with one of which sample
(a)Fall into the fiducial confidence ellipse of one of which sample
(b)Without falling into the fiducial confidence ellipse of any group of sample.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710483442.XA CN107703173A (en) | 2017-06-23 | 2017-06-23 | A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710483442.XA CN107703173A (en) | 2017-06-23 | 2017-06-23 | A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107703173A true CN107703173A (en) | 2018-02-16 |
Family
ID=61170449
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710483442.XA Pending CN107703173A (en) | 2017-06-23 | 2017-06-23 | A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107703173A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109374667A (en) * | 2018-11-05 | 2019-02-22 | 中国农业科学院农业质量标准与检测技术研究所 | A kind of NMR spectrum method identifying ultra-high-temperature sterilized milk and reconstituted milk |
CN110751198A (en) * | 2019-10-15 | 2020-02-04 | 南京航空航天大学 | Wood type identification system and method based on RFID (radio frequency identification) tag |
CN111948329A (en) * | 2020-06-01 | 2020-11-17 | 上海市质量监督检验技术研究院 | Method for identifying pteroceltis tatarinowii and pteroceltis tenuipes |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002056000A1 (en) * | 2001-01-12 | 2002-07-18 | Universite De Victor Segalen Bordeaux 2 | Discriminating method with location and/or identification of situations of biological perturbations by spectrometry and pattern recognition |
SI23011A (en) * | 2009-03-26 | 2010-09-30 | Institut@@quot@JoĹľef@Stefan@quot | Method and device for instantaneous determination of moisture in wood |
CN102269720A (en) * | 2011-04-28 | 2011-12-07 | 王士桥 | Method for identifying edible oil quality through combination of nuclear magnetic resonance and pattern recognition method |
CN103728331A (en) * | 2014-01-24 | 2014-04-16 | 西南民族大学 | Detection method for pulse-activating injection |
CN105548233A (en) * | 2015-10-30 | 2016-05-04 | 中国科学院武汉物理与数学研究所 | Method for discriminating acacia honey and rape honey on basis of H-nuclear magnetic resonance |
CN106124551A (en) * | 2016-08-26 | 2016-11-16 | 浙江师范大学 | Fructus Citri Sarcodactylis hydrogen nuclear magnetic resonance finger printing and method for building up thereof and purposes |
-
2017
- 2017-06-23 CN CN201710483442.XA patent/CN107703173A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002056000A1 (en) * | 2001-01-12 | 2002-07-18 | Universite De Victor Segalen Bordeaux 2 | Discriminating method with location and/or identification of situations of biological perturbations by spectrometry and pattern recognition |
SI23011A (en) * | 2009-03-26 | 2010-09-30 | Institut@@quot@JoĹľef@Stefan@quot | Method and device for instantaneous determination of moisture in wood |
CN102269720A (en) * | 2011-04-28 | 2011-12-07 | 王士桥 | Method for identifying edible oil quality through combination of nuclear magnetic resonance and pattern recognition method |
CN103728331A (en) * | 2014-01-24 | 2014-04-16 | 西南民族大学 | Detection method for pulse-activating injection |
CN105548233A (en) * | 2015-10-30 | 2016-05-04 | 中国科学院武汉物理与数学研究所 | Method for discriminating acacia honey and rape honey on basis of H-nuclear magnetic resonance |
CN106124551A (en) * | 2016-08-26 | 2016-11-16 | 浙江师范大学 | Fructus Citri Sarcodactylis hydrogen nuclear magnetic resonance finger printing and method for building up thereof and purposes |
Non-Patent Citations (2)
Title |
---|
郑起帆 等: "基于 1H-NMR的四个茶山普洱生茶代谢组学研究", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》 * |
马红梅 等: "《实用药物研发仪器分析》", 30 September 2014 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109374667A (en) * | 2018-11-05 | 2019-02-22 | 中国农业科学院农业质量标准与检测技术研究所 | A kind of NMR spectrum method identifying ultra-high-temperature sterilized milk and reconstituted milk |
CN109374667B (en) * | 2018-11-05 | 2021-09-21 | 中国农业科学院农业质量标准与检测技术研究所 | Nuclear magnetic resonance spectrum method for identifying ultra-high temperature sterilized milk and reconstituted milk |
CN110751198A (en) * | 2019-10-15 | 2020-02-04 | 南京航空航天大学 | Wood type identification system and method based on RFID (radio frequency identification) tag |
CN110751198B (en) * | 2019-10-15 | 2022-04-15 | 南京航空航天大学 | Wood type identification system and method based on RFID (radio frequency identification) tag |
CN111948329A (en) * | 2020-06-01 | 2020-11-17 | 上海市质量监督检验技术研究院 | Method for identifying pteroceltis tatarinowii and pteroceltis tenuipes |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Laghi et al. | Nuclear magnetic resonance for foodomics beyond food analysis | |
JP4710012B2 (en) | Visible light / near infrared spectroscopic analysis method and apparatus thereof | |
Barding et al. | Quantitative NMR for bioanalysis and metabolomics | |
Verpoorte et al. | Ethnopharmacology and systems biology: a perfect holistic match | |
CN105510372B (en) | Set up the true and false model method of DPLS BS UVE quick discriminating honey | |
Wang et al. | Fully-automated high-throughput NMR system for screening of haploid kernels of maize (corn) by measurement of oil content | |
CN103134850B (en) | A kind of tea leaf quality method for quick based on characteristic perfume | |
CN106950241B (en) | A kind of method of other adulterated oil type and contents in prediction tea oil | |
CN108629365A (en) | Analyze data analysis device and analysis data analysis method | |
CN107703173A (en) | A kind of identification apparatus and method of the rare species timber based on 1H NMR datas storehouse | |
Jie et al. | Review on the recent progress of non-destructive detection technology for internal quality of watermelon | |
CN101769910A (en) | Method for screening malignant ovarian tumor markers from blood serum metabolic profiling | |
CN109791124A (en) | Analytical data of mass spectrum resolver and analytic method | |
CN106918572B (en) | The assay method of potato content in potato compounding staple food | |
CN109669023A (en) | A kind of soil attribute prediction technique based on Multi-sensor Fusion | |
CN106841083A (en) | Sesame oil quality detecting method based on near-infrared spectrum technique | |
CN108195790A (en) | The method that chevon quality detection database is established based near infrared spectrum | |
CN107505346A (en) | Predict the method that the chemical displacement value of NMR spin systems in biological fluid sample is particularly in class of fluids sample | |
Wiklund | Multivariate data analysis for Omics | |
KR100934410B1 (en) | Simple determination of seed weights in crops using near infrared reflectance spectroscopy | |
CN106018452A (en) | Peanut variety nondestructive testing method based on nuclear magnetic resonance technology | |
CN109187481A (en) | One kind being based on Fe3O4The Pesticides Testing method of@Au NPs and molecular engram | |
US6895340B2 (en) | Method of molecular structure recognition | |
JP6281973B2 (en) | Method for expressing properties of a mixture sample, method for evaluating properties of a mixture sample, method for identifying attributes of a mixture sample, and method for processing an electromagnetic wave signal derived from a mixture sample | |
CN109001143A (en) | A kind of mid-infrared light spectrometry of sensitive prediction Chinese ephedra quality characteristic |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180216 |