NIRS Consortium Round-Robin

Starch Determination  

 

were sent out to labs on April 14th and we expect analysis and reports to be completed within 14 days of a lab’s receipt of samples.  Data will be compiled by Mary Beth Hall.  After this, samples will be identified and results discussed with the labs.  As noted at the conference, the analyzed data from labs, with lab anonymity preserved, is likely to be published by Mary Beth Hall in a research journal.

 
 

     Stemming from discussion at the NIRS Consortium annual conference in February, 14 member labs have agreed to participate in a Starch Sample Exchange, with 3 more members interested in participating in the equation development in some capacity.  Mary Beth Hall and Don Sapienza have developed a standardized protocol for labs to follow.  Labs will analyze 11 samples (those labs participating in the

 

NFTA program, will additionally analyze their March corn silage sample) prepared by Mary Beth Hall.  All subsamples have been scanned on the Consortium master instrument to record spectra for homogeneity.  Each lab will analyze the samples using their “routine” method and report information about their instrument, reagents, and results of starch and sugar.  

     The unknown samples

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Text Box: FOSS Software Package Deadline May 2, 2003 A reminder to those labs considering a software upgrade from FOSS North America: The deadline for their promotional package outlined at the annual conference is TODAY, May 2, 2003. If your lab has any questions regarding this package, contact Mark Host, Regional Sales Manager, ph: 952-974-9892 mhost@fossnorthamerica.com
 

 

 

 

 

 

 

 


Microwave versus Oven Drying

Under Study at USDFRC

- Patty Laskowski, Susan Selman

 

After the Consortium presentation on microwave versus oven drying at the annual conference, along with Don Sapienza’s image of starch granules having undergone different drying methods, the group had a lot to think about in terms of where to go.  After some work from Dave Taysom, Don Meyer, and Mark Heidgerken in developing a description of issues, the project has been taken on by the US Dairy Forage Research Center. 

Last fall, the USDFRC sent out detailed surveys asking labs using microwave ovens to report on sample handling and drying using microwaves as well as details of the oven itself.

Four labs have decided to participate in this study.  They will analyze 3 different

 
 
Text Box: Plan to Attend the Summer NIRS Workshop! The NIRS Consortium will again put on a summer NIRS workshop. This year it will be held in the month of August. We will keep you posted on the details as they develop. Plan on getting in some practice with the FOSS ISIscan software! Send in your comments if you have any ideas on what you’d like to see in training at the Workshop.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Release of NIRS Consortium 5000 Equations  

- Susan Selman

 

Following are Standard Error of Prediction Stats for 5000 equations for certain constituents.

 

Standard Error of Prediction Stats for 5000 hay equations for certain constituents.

5000 Hay equations

 

 

 

 

 

 

 

 

 

MEAN

MEAN

 

 

 

 

 

SEP

LAB

NIR

BIAS

R2

SLOPE

#

PROTEIN

1.32

18.58

18.56

0.025

0.95

1.04

43

ADF

1.67

32.94

33.38

-0.433

0.94

1.10

40

NDF

2.34

43.60

43.40

0.176

0.88

0.99

23

NDF

1.84

43.38

43.65

-0.277

0.94

1.07

29

dNDF

2.25

20.01

19.9

0.103

0.86

1.08

19

Ash

0.617

9.67

9.73

-0.062

0.52

0.73

40

Lignin

0.629

7.01

7.15

0.152

0.70

1.18

8

I have two SEP for NDF.  I have used two different sources of spectra for validation for NDF.  The first one is a mixture of data from 3 labs, that has been submitted for next years updates.  The last NDF are samples that went in for updates for this year, not used in the calibration.  # number of samples tested on the equation.

 

Standard Error of Prediction Stats for 5000 haylage equations for certain constituents.

                                                                       

5000 

Haylage

Equations

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

MEAN

MEAN

 

 

 

 

 

SEP

LAB

NIR

BIAS

R2

SLOPE

#

PROTEIN

1.07

20.81

20.58

0.228

0.94

1.08

18

ADF

2.92

32.26

34.00

-1.75

0.88

1.06

25

NDF

2.57

45.40

45.61

-0.211

0.91

0.90

26

NDF

2.38

42.92

42.71

0.202

0.91

0.94

22

dNDF

1.85

21.62

21.51

0.109

0.95

0.85

13

Ash

1.37

10.98

10.76

0.220

0.87

1.06

32

Lignin

0.83

7.49

6.67

0.821

0.32

0.71

3

 

I have two sep for NDF.  I have used two different sources of spectra for validation for NDF.  The first one is a mixture of data from 3 labs, that has been submitted for next years updates.  The last NDF are samples that went in for updates for this year, not used in the calibration.   # number of samples tested on the equation.  Lignin low R2 is on only 3 samples.

 

5000 Grass equations

 

 

 

 

 

 

 

MEAN

MEAN

 

 

 

 

 

SEP

LAB

NIR

BIAS

R2

SLOPE

          #

PROTEIN

0.943

12.62

12.72

-0.102

0.95

1.05

           15

ADF

2.34

36.68

35.75

0.93

0.89

0.99

           15

NDF

2.83

55.34

54.97

0.367

0.86

0.99

           12

dNDF

2.33

27.84

27.74

0.098

0.57

0.90

           11

Ash

1.50

10.39

9.76

0.626

0.78

1.45

           13

Lignin

Not

Enough

Sample

To validate

 

 

 

 

 

 

 

Standard Error of Predictions on Past NFTA Samples

 

Past NFTA Predictions

5000 Legume Hay equation

As Received

 

Mean

Mean

 

 

GH=.469     NH=.211

 

Lhay NFTA

SEP

Lab

NIRS

Bias

R2

Slope

 

Protein

0.855

19.25

20.01

-0.764

0.97

0.97

 

ADF

1.16

28.83

29.35

-0.521

0.96

1.1

 

NDF

1.31

35.11

36.05

-0.942

0.97

1.04