Slide 1
Hello. My name is Sarah Cervantes-Pahm, and today, I would like to share
with you some of the results of our project entitled, "Evaluation of
in-vitro procedures to predict the digestible energy (abbreviated as DE) in
yellow dent corn."

Slide 2
As a background, out of the 600-plus corn samples that the company had, we
were able to obtain 50 corn samples from Pioneer Hi-Bred International,
Inc., a seed company based in Iowa. These 50 corn samples were chosen to
provide variability to the database. And the variability was based on
chemical characteristics, the concentration of DE, and apparent total tract
digestibility (abbreviated as ATTD) of GE, or gross energy, in yellow dent
corn.

Slide 3
This table shows some of the chemical characteristics of the corn samples.
Among the chemical characteristics that were analyzed, the concentration of
crude fat and NDF have the greatest coefficient of variation. From the
table, we see that the average concentration of crude fat is 4.5% with a
range of 3.11 to 7.57%, with a coefficient of variation of 30.16%. We could
also see that the average concentration of NDF is 2.66%, ranging between
1.09 to 7.56%, with a coefficient of variation of 55.16%. Based on this
table, we could also see that despite the high coefficient of variation of
crude fat and NDF, the concentration of GE, DE, and the in vivo ATTD of GE
was not very wide. Here we see that the mean, or the average, GE among the
50 corn samples was 4,513 kcal/kg dry matter, with a range between 4,410 and
4,720 kcal/kg dry matter, with a coefficient of variation of only 1.78%.
Likewise, we would see for the DE that the average DE of the 50 corn samples
was 4,040 kcal, ranging between 3,777 and 4,267 kcal/kg dry matter. And the
CV was only at 2.6%. In vivo ATTD of GE for the 50 corn samples averaged
89.52%, with a range of 85.12 to 92.32%, with a CV of only 1.68%. These
small variations in DE and ATTD of GE tells us that we need a procedure that
is sufficiently sensitive to detect the small differences in the
concentration of DE or the ATTD in these corn samples.

Slide 4
So to go to our objectives, the first objective of this experiment was to
evaluate in vitro procedures and techniques that could correlate in vitro
dry matter disappearance with the concentration of DE in these 50 corn
samples. The second objective of the experiment was to develop a regression
equation that can predict the concentration of DE in corn.

Slide 5
And the third objective was to evaluate the suitability of the Daisy II
incubator as an alternative to the traditional water bath. And this is a
picture of the Daisy incubator that we used for this experiment.

Slide 6
For this experiment, we considered the Boisen and Fernandez procedure as the
standard procedure. And this procedure was published in 1997. The Boisen and
Fernandez procedure is a three-step in vitro procedure that simulates the
digestion processes in vivo. Step 1 made use of pepsin to simulate gastric
digestion, and in this procedure, pepsin was added at 25 mg/ml and the
samples were incubated for two hours. Step 2 made use of pancreatin to
simulate small intestinal digestion, and pancreatin was added at 100 mg and
the samples were incubated for four hours. For Step 3, viscozyme was added
at 5 ml, and viscozyme was added to simulate hindgut fermentation, and the
samples were incubated for 18 hours. So in total, using this procedure, the
incubation period consisted of 24 hours. We obtained in vitro dry matter
disappearance values for all of the 50 corn samples using this procedure.
And as you can see, the relationship between in vitro dry matter
disappearance and the DE of the 50 corn samples was not very good because we
only got an R-squared of 0.12. Therefore, we decided to conduct additional
experiments to come up with in vitro procedures that would correlated well
with the concentration of DE in corn. 

Slide 7
These experiments that we did include looking at the effect of sample size
on in vitro dry matter disappearance; we looked at samples sizes ranging
from 0.5 grams to 3.5 grams. We also looked at the effect of the type of
incubation on in vitro dry matter disappearance, and we compared the Daisy
incubator with the water bath. We also looked at the effect of length of
incubation on in vitro dry matter disappearance. And lastly we evaluated the
three in vitro procedures to determine the best in vitro procedure for our
purpose.

Slide 8
The three in vitro procedures that we evaluated were the Boisen and
Fernandez procedure, published in 1997, the procedure of Huang and others
published in 2003, and we also modified the Boisen and Fernandez procedure.
These three procedures have some similarities and some differences. All of
these procedures are three-step procedures, and they use the same enzyme in
the first two steps in the procedure. However, whereas the Boisen and
Fernandez procedure requires the use of pepsin or the inclusion of pepsin at
25 mg, and the samples are incubated for two hours, the procedure of Huang
and others includes pepsin at only 10 mg/ml, but the samples are incubated
for six hours. Likewise, whereas Boisen and Fernandez only includes 100 mg
of pancreatin, and only incubates the samples for four hours, the procedure
of Huang and others in 2003 requires an 18-hour incubation of the samples
and pancreatin is only added at 50 mg/ml. On the third step of the
procedure, Boisen and Fernandez uses viscozyme to simulate hindgut
fermentation, and viscozyme is added at 5 ml, and the samples are incubated
for 18 hours. Whereas for the procedure of Huang and others, for the last
step of the procedures they make use of cellulase, and they add cellulase at
20 ml, and the samples are incubated for 24 hours. So in total, the Boisen
and Fernandez procedure incubates the samples for 24 hours, whereas the
procedure of Huang and others incubates the samples for 48 hours. This is
the basis for the study of the effect of length of incubation later on. For
the third procedure, we basically modified the Boisen and Fernandez
procedure only on the third step of the procedure, where instead of adding
viscozyme, we added the fecal inoculum. 

Slide 9
Now let's look at the results.

Slide 10
This graph shows the effect of sample size on in vitro dry matter
disappearance and in vitro gross energy disappearance (abbreviated as
IVGED). In the x-axis, you see the sample weight, which ranged between 0.5
to 3.5 grams, and on the y-axis, you see the in vitro dry matter
digestibility or in vitro dry matter disappearance. The blue line on top
represents in vitro gross energy disappearance using the Daisy incubator,
and the red line on top represents in vitro dry matter disappearance using
the Daisy incubator. The last two lines at the bottom use the water bath,
and the green line represents in vitro dry matter disappearance, and the
orange line represents in vitro gross energy disappearance. And you will see
immediately that we see a quadratic reduction in in vitro dry matter
disappearance and in vitro gross energy disappearance when we increase the
sample weight from 0.5 to 3.5 grams. And this observation is true for both
types of incubator, whether we use the Daisy incubator or the water bath.
This graph tells us that we could not increase the sample weight to more
than 0.5 grams. 

Slide 11
This graph present the effect of type of incubation and length of incubation
on in vitro dry matter disappearance. The first two bars on your left
represent in vitro dry matter disappearance using the water bath, and the
two bars on your right represent in vitro dry matter disappearance using the
Daisy incubator. The blue bar represents short incubation, which is the
24-hour incubation, and the yellow bar represents the long incubation, which
is the 48-hour incubation. And in this procedure, we all used the viscozyme.
First, we could see that regardless of the type of incubation, the in vitro
dry matter disappearance of samples incubated for the long period, or the 48
hours, has greater in vitro dry matter disappearance than corn samples that
were incubated for a short time, or for 24 hours. Second, we could also
observe a type and length interaction. And we could see this if we look at
the difference between the in vitro dry matter disappearance of corn samples
incubated for a short and long period using the water bath, which was
twofold greater than the difference in in vitro dry matter disappearance of
corn samples incubated for the same period using the Daisy incubator. Third,
we could also see that the in vitro dry matter disappearance of corn samples
incubated using the water bath for a long period is not different from the
in vitro dry matter disappearance of corn samples that was incubated for a
long period using the Daisy incubator.

Slide 12
This table shows the regression equations to predict the concentration of DE
derived from in vitro dry matter disappearance. On the first column is the
type of incubator and that is the water bath or the Daisy incubator. The
second column, you will see the length of incubation, which is short
incubation or 24-hour incubation or the long incubation which is the 48-hour
incubation. And in the third column you would see the regression equations.
On the fourth column, you would see the r-squared or the relationship
between in vitro dry matter disappearance and DE of the corn samples. And on
the last column, you would see the p value. As we could see in this table,
whether we used the water bath or the Daisy incubator, we observed that
increasing the length of incubation improves the r-squared of the regression
equation. And here, you see the water bath, we could see that for the short
incubation, r-squared is only at 0.03; however, when we increased the
incubation time to 48 hours, the r-squared improved to 0.4 with a p value of
0.13. In the same way, we observe that for the short incubation, the
r-squared using the Daisy incubator is only at 0.11; however, when we
increase the incubation time to 48 hours from 24 hours, we also improve the
r-squared to 0.72 with a p value of 0.02. Because we obtain a better
r-squared between in vitro dry matter disappearance and the concentration of
DE in the corn samples when we incubate the samples for a long time using
the Daisy incubator, we decided to use the Daisy incubator for our next
experiment.

Slide 13
This table presents the regression equations to predict the DE in corn using
in vitro dry matter disappearance from three in vitro procedures. As
mentioned previously, the three in vitro procedures that were evaluated were
the use of viscozyme, the use of cellulase, and the use of the fecal
inoculum, which you could find in the first column of this table. The second
column shows the length of incubation Ð that's the short or the long
incubation. The third column shows the regression equations derived from
these in vitro procedures. The fourth column shows the r-squared, and the
last column the p value. And once again, we see the same pattern as was
observed in the previous experiment where, when we increase the incubation
time from 24 to 48 hours, we observe an improvement in the r-squared, and
this is true for viscozyme and cellulase. And viscozyme, we see that for the
short incubation, r-squared was only at 0.1, but when we increased the
incubation time from 24 to 48 hours, r-squared improved from 0.1 to 0.53%,
and we have a p value of 0.06. In the same way, cellulase, for the short
incubation, only provided an r-squared of 0.06, but when we increase the
incubation time from 24 to 48 hours, we see an improvement in the r-squared
from 0.06 to 0.41%, with a p value of 0.12. However, the improvement in the
r-squared when we increased incubation time from short to long was not
observed when we used the fecal inoculum. Based on these results, we were
able to come up with a final in vitro procedure.

Slide 14
The final in vitro procedure makes use of 0.5 grams of sample, requires a
long incubation time, which is a 48-hour incubation time, requires the use
of the Daisy incubator, and we basically follow the Boisen and Fernandez
procedure. And based on this final in vitro procedure, we were able to come
up with a regression equation, as presented, that is able to explain 53% of
the variability in the DE in these corn samples.

Slide 15
So based on this final in vitro procedure, we obtained the in vitro dry
matter disappearance of 28 corn samples which we used to build a regression
model. We also included some chemical characteristics of the corn samples to
improve the regression model. So the final regression model to predict the
DE in the corn samples includes five components. And those five components
are: the in vitro dry matter disappearance, NDF, gross energy, starch, and
corn density. And with the addition of the chemical characteristics, we were
able to improve the r-squared of the regression equation from the original
0.53 to 0.81 with a p value of 0.001. But in order for us to know if we
could use this regression equation to other corn samples which was not
included in building the model, we decided to validate this regression
equation by obtaining ten corn samples that were independent of the 28 corn
samples that we used for model building. So we determined the in vitro dry
matter disappearance of these corn samples using the final in vitro
procedure, and fed the in vitro dry matter disappearance values, as well as
the physical and chemical characteristics of the ten corn samples, into the
model to come up with a predicted DE. 

Slide 16
This graph shows the relationship between in vivo DE in the x-axis and the
predicted DE in the y-axis. As you could see, some of the red dots, which
represented the corn samples, were on the line and some were quite far from
the line, but in general we were able to obtain an r-squared of 0.72. The
slope of the line is also not different from zero, and we validated the
regression model by using the mean square prediction of error, which looks
at the difference between the in vivo DE and the predicted DE. The mean
square prediction of error is equal to 0.007, and you could see that in the
upper right hand. And the mean square prediction of error of 0.007 is less
than three times the mean square error of the regression equation. This
tells us that we were able to come up with a regression model that is able
to predict the DE in corn with reasonable accuracy. 

Slide 17
In conclusion, an in vitro procedure was established that was able to
explain 53% of the variability in DE in corn. And this procedure makes use
of the Daisy incubator, requires a 48-hour incubation period, and makes use
of the Boisen and Fernandez procedure that was published in 1997. And in
this experiment, a validated prediction equation to determine the DE in corn
with reasonable accuracy was developed.

Slide 18
Thank you for your attention.

Slide 19
If you are interested with more nutrition facts, please visit our website,
which is found at the address on your screen.