Brookings, South Dakota
May, 2009
Source: SGS Seed
and Crop Services May Newsletter
by
Kalyn Brix-Davis, Genetic Services
Director, Branch Manager
With seed testing that may require
individual seed analysis for specific trait or events and/or a
shortage of seed available for testing in situations such as
elite inbred seed, sequential testing is a tool that can manage
risk, short seed stocks and expense of testing. Sequential
testing is a process in which a testing plan is developed based
on statistics on the amount of risk associated with testing a
minimum amount of seed at the initial testing period. The level
of rejection is higher than normal to allow for a lower amount
to be tested at the initial period, while a secondary tier of
testing is established when this level of rejection is reached.
In the overall scheme, the amount of primary and secondary
testing required is less than what would be required with the
lower level of rejection required with testing more seeds during
the initial testing period.
SeedCalc7 was developed by seed quality statistician to assist
in calculation of purity/impurity levels in seed lots with the
ability to set the confidence level and determine the amount of
risk that is associated with different sampling/testing
schemes. This Microsoft Excel program is available at the
International Seed Testing Association Statistics committee
website (http://seedtest.org/en/content---1--1143.html).
In the table below are various number of seeds that are tested
in genetic or trait purity testing (growouts, electrophoresis,
insecticide trait) with different number of deviants for the
testing and the calculated purity level and the upper bound
limit at a 95% level. In our example, a 95% confidence level
that means that the purity level will be below the upper bound
limit 95% of the time.
|
Number
of Seeds |
Number
of deviants |
Calculated Purity % |
Upper
Bound Limit |
|
50 |
0 |
100 |
5.82 |
|
100 |
0 |
100 |
2.95 |
|
50 |
1 |
98 |
9.14 |
|
100 |
1 |
99 |
4.66 |
|
100 |
2 |
98 |
6.16 |
|
100 |
5 |
95 |
10.23 |
|
90 |
0 |
100 |
3.27 |
|
90 |
3 |
96.7 |
8.39 |
|
180 |
7 |
96.1 |
7.18 |
For example, if
50 seed electrophoresis test resulted with 0 offtypes or selfs,
the calculated purity level would be 100% with a upper bound
level of 5.82% impurity. With 1 deviant in 50 seeds, the purity
level is 98% with an upper bound level of 9.14% impurity. If an
additional 50 seeds are tested and 0 deviants (total tested =
100, 1 deviant), the purity level is 99% with an upper bound
level of 4.66% impurity.
By using sequential sampling, a testing plan can be devised to
maximize information and costs while testing genetic purity to a
confidence level and quality standard that your company is
comfortable in providing to your customer. If you would like
additional information, please contact Kalyn at
kalyn.brixdavis@sgs.com. |
