Method of predicting soybean yield

Patent metadata

Field	Value
Publication number	US-12073350-B2
Application number	US-202017603731-A
Country	US
Kind code	B2
Filing date	Apr 16, 2020
Priority date	Apr 16, 2019
Publication date	Aug 27, 2024
Grant date	Aug 27, 2024

Abstract

To provide a method for predicting a soybean yield at an early stage with high accuracy.The method for predicting a soybean yield comprises: acquiring analytical data of one or more components from a leaf sample collected from the soybean; and predicting a soybean yield using a correlation between the data and a soybean yield.

First claim

The invention claimed is:

1. A method, the method comprising:
choosing a plurality of components of a soybean plant;
acquiring first analytical data of each of the plurality of chosen components of a leaf sample collected from a plurality of first soybean plants;
constructing, by a processor, a first yield prediction model in a form of a machine learning model using the first analytical data of each of the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants;
calculating a variable importance in projection (VIP) value for each of the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants with respect to the first yield prediction model;
selecting a subset of the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants based upon the VIP values, the subset of the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants including fewer components than the plurality of chosen components;
constructing, by the processor, a second yield prediction model in a form of a machine learning model using the first analytical data of each of the components in the subset of the plurality of chosen components which includes fewer components than the plurality of chosen components;
acquiring second analytical data of the components of the subset of the plurality of chosen components which includes fewer components than the plurality of chosen components from a leaf sample collected from a second soybean plant from a field;
predicting a soybean yield using the second analytical data and the second yield prediction model, which was constructed using the first analytical data of each of the components in the subset of the plurality of chosen components which includes fewer components than the plurality of chosen components;
selecting a material for the field from which the second soybean plant came based upon the predicted soybean yield;
comparing an actual soybean yield of the second soybean plant and the predicted soybean yield; and
optimizing the second yield prediction model based upon a result of the comparison of the actual soybean yield and the predicted soybean yield.

2. The method according to claim 1 , wherein the first analytical data is corrected by a pooled QC method.

3. The method according to claim 1 , wherein the first analytical data is corrected by an internal standard substance.

4. The method according to claim 1 , wherein the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants used for constructing the first yield prediction model are at least two components selected from the group consisting of components having an accurate mass (m/z), provided by mass analysis, of 139 to 1,156.

5. The method according to claim 1 , wherein the plurality of chosen components of the leaf sample collected from the plurality of first soybean plants used for constructing the first yield prediction model are at least two components selected from the group consisting of components described in the following Tables A1a to 1c, defined by an accurate mass (m/z) provided by mass analysis:

TABLE A1a

Component

Component

No.
m/z
No.
m/z

1
139.0389
81
259.0827

2
141.9592
82
259.2076

3
147.0435
83
261.1501

4
147.0446
84
261.2233

5
149.0234
85
263.2381

6
149.0241
86
264.2335

7
161.0606
87
265.1440

8
163.0398
88
269.0818

9
163.1325
89
271.0618

10
165.0550
90
271.0619

11
170.0974
91
271.2280

12
171.1501
92
273.0769

13
175.1486
93
274.0541

14
177.0551
94
274.0928

15
179.0717
95
274.1606

16
181.1232
96
275.2020

17
181.1237
97
275.2023

18
183.1865
98
277.2184

19
186.0921
99
277.2186

20
189.1278
100
277.2186

21
190.0506
101
279.0512

22
191.1437
102
279.0515

23
191.1439
103
279.0951

24
193.0859
104
279.1610

25
193.0861
105
279.1611

26
193.1597
106
279.2320

27
194.1182
107
279.23

Assignees

• Kao Corp

Inventors

• Fujimatsu Teruhisa
• Suetsugu Mai
• Deguchi Jun
• Endo Keiji

Classifications

• G06Q50/02 Primary

  Agriculture; Fishing; Forestry; Mining · CPC title

• G06Q10/06395

  Quality analysis or management · CPC title

• G06Q10/06375 Primary

  Prediction of business process outcome or impact based on a proposed change · CPC title

• A01G7/00

  Botany in general · CPC title