1.3 assessment of disease intensity introduction and definitions methods for assessing disease...

1.3 Assessment of disease intensity

•Introduction and definitions

•Methods for assessing disease intensity

•Methods for analyzing disease records

•Methods for presenting disease records

•Concluding remarks

Why do we need (want) to assess disease intensity?or

What are the uses of disease intensity records

•For making decision concerning disease management

•For quantifying the effect of the pathogens on the host plant and on yield

•For determining the efficacy of various control measures (pesticides, genotype resistance, agronomic measures, etc.)

•For verifying the quality of agronomic products and seed

Methods to assess disease intensity

Disease incidence:

The proportion of infected host units, out of the total units sampled

4/10*100=40%

4/10*100=40%

Methods to assess disease intensity

Disease severity:

The proportion of infected part (or the infected area), in each of the sampled units

(60+10+30+1)/10*100=10.1%

60 10 30

1 0 0 0 0 0 0

Percentage scale (0-100%)

4 2 2 1 0

0 0 0 0 0

(4+2+2+1)/10*100=0.9

Index scale (0-4)

What is better: to assess disease incidence

or to assess disease severity?

•What is easier?

•What is less subjective to errors?

•What represents better the effect of the pathogen on the host?

Can disease incidence records be converted to disease severity ?

100Disease incidence (%)

0

Dis

ease

sev

erit

y (%

)

0

100

50

50

Disease incidence (%)0 100

Dis

ease

sev

erit

y (%

)

0

10

50

5

What scale to use in the assessments?

•Number of infected blossoms in a blossom-cluster

•Number of infected blossoms in a blossom-cluster, out of the total number of blossoms in that cluster

•Number of infected blossom-clusters in a tree

•Number of infected blossom-clusters in a tree, out of the total number of blossom-clusters in that tree

Sampling unit: a tree

Sampling unit: blossom-cluster

•The number of trees with infected blossom-clusters

•Number of trees with infected blossom-clusters out of the number of trees in the orchard

Sampling unit: an orchard

Which part of the plant should be assessed?

Flag leafLeaf -1

Leaf -3

Leaf -5

Leaf -2

Leaf -4

Septoria leaf blotch caused by Mycosphaerella graminicola

Which type of symptom should be assessed?

Fruit rot Fruit ghost spots

Grey mould caused by Botrytis cinerea

Leaf infectionStem infection

How to assess disease severity?

The Webber-Fechner low:

1. The eye can detect logarithm differences

1% 3%

40% 43%

How to assess disease severity?

The Webber-Fechner low:

2. Up to 50% severity - the eye detect the diseased, out of the healthy portionabove 50% severity - the eye detect the healthy, out of the diseased portion

3% 12%

88% 97%

Scales for assessment of disease severity

The Horsfall and Barratt scale

1) 0%

2) 1 - 2%

3) 2 - 6%

4) 6 - 12%

5) 12 - 25%

6) 25 - 50%

12) 100%

11) 98 - 99%

10) 94 - 98%

9) 88 - 94%

8) 75 - 88%

7) 50 - 75%

Pictorial scale (% diseased area)

0% 1% 2% 5% 10%

20% 50% 80% 90% 95%

Scales for assessment of disease severity

Descriptive verbal scale (%)

0% Late blight is not seen in the field

0.5% Up to one lesion per 100 m2

1% Up to 5 lesions per plant

5% One out of 10 leaflets is infected

25% Almost all leaflets are infected. The field still looks green, but with many infected foci.

50% Half of the foliage is infected. The field looks infected.

75% Most of the foliage is diseased. The field looks defoliated with some green points.

95% Leaves were defoliated. Some of the stems are still green.

Scales for assessment of disease severity

Descriptive indexes

Number of lesions per leaf

0 = no lesions

1 = 1-10 lesions

2 = 10-20 lesions

3 = 20-50 lesions

4 = >50 lesions

Index 0 4

0 = Healthy

1 = Low disease severity

2 = Moderate disease severity

3 = High disease severity

4 = Dead plant

Scales for assessment of disease severity

Descriptive indexes

Number of lesions per leaf

0 = no lesions

5 = 1-10 lesions

15 = 10-20 lesions

40 = 30-50 lesions

60 = 50-70 lesions

85 = 70-100 lesions

Scales for assessment of disease severity

Number of lesions per leaf

0 = no lesions

1 = 1-10 lesions

2 = 10-20 lesions

3 = 30-50 lesions

4 =50-70 lesions

5 = 70-100 lesions

Descriptive index - sporulation of Botrytis cinerea on tomato stem pieces

1 2 3 4 5

Scales for assessment of disease severity

Methods for analyzing disease records - Descriptive statistics

Sample: 5,100,2,3,6,5,5,5,6,4

Frequent record: the most frequent record in the sample 5.0

Mean: sum of the values divided by the number of records 14.1

Mid-range: mean of the two extreme values 51.0

Median: the value that split the distribution of the records 5.0

Nominal variableidentification of the records

Methods for analyzing disease records - Variable types

numbers on the shirts of football players

(1,2,7,12)

6

school grade

(grade A, B, C….F)

Methods for analyzing disease records - Variable types

Ordinal variable

identification of the records

the order of the records

Methods for analyzing disease records - Variable types

weight, length….20 kg

5 kg5 kg

5 kg5 kg

Continuous variable

identification of the records

the order of the records

the ratio between records

Methods for analyzing disease records

Statistics and variable types

Variable type

statisticsFrequent mid-range median meannumber

continuous

nominal

ordinal

Methods for analyzing disease records

Statistics and variable types

The outcome is that calculation of means and conducting analysis of variance (ANOVA) tests may be done

only

for disease records that are continuous variables

The Horsfall and Barratt scale

Type of variable:

ordinal

Most frequent no.

Mid-range

Median

Mean

1) 0%

2) 1 - 2%

3) 2 - 6%

4) 6 - 12%

5) 12 - 25%

6) 25 - 50%

12) 100%

11) 98 - 99%

10) 94 - 98%

9) 88 - 94%

8) 75 - 88%

7) 50 - 75%

Pictorial scale

0%

1%

2%

5%

10%

20% 50% 80% 90% 95%

Type of variable:

continuous

Most frequent no.

Mid-range

Median

Mean

Descriptive verbal scale

0% Late blight is not seen in the field

0.5% Few infected plants. Up to one lesion per 100 sq. meters

1% Up to 5 lesions per plant

5% One out of 10 leaflets is infected

25% Almost all leaflets are infected. The field still looks green, but with many infected foci.

50% Half of the foliage is infected. The field looks infected.

75% Most of the foliage is diseased. The field looks defoliated with some green points.

95% Leaves were defoliated. Some of the stems are still green.

Type of variable:

continuous

Most frequent no.

Mid-range

Median

Mean

Type of variable:

ordinal

Most frequent no.

Mid-range

Median

Mean

Number of lesions per leaf

0 = no lesions

1 = 1-10 lesions

2 = 10-20 lesions

3 = 20-50 lesions

4 = >50 lesions

Index 0 4

0 = Healthy

1 = Low disease severity

2 = Moderate disease severity

3 = High disease severity

4 = Dead plant

Type of variable:

ordinal

Most frequent no.

Mid-range

Median

Mean

Number of lesions per leaf

0 = no lesions

1 = 1-10 lesions

2 = 10-20 lesions

3 = 30-50 lesions

4 =50-70 lesions

5 = 70-100 lesions

Type of variable:

continuous

Most frequent no.

Mid-range

Median

Mean

Number of lesions per leaf

0 = no lesions

5 = 1-10 lesions

15 = 10-20 lesions

40 = 30-50 lesions

60 = 50-70 lesions

85 = 70-100 lesions

Descriptive index - sporulation of Botrytis cinerea on tomato step pieces

1 2 3 4 5Type of variable:

ordinal

Most frequent no.

Mid-range

Median

Mean

Descriptive index - sporulation of Botrytis cinerea on tomato stem pieces

0 1 2 3 4

Treatment 1: 4 replicates: 1,2,2,1

Mean index = 1.5

Treatment 2: 4 replicates: 3,3,4,4

Mean index = 3.5

0

5

10

15

20

0 1 2 3 4

Sporulation index

No.

of

spor

es (

x106 )

Mean number of spores: 1,500,000

Mean number of spores: 10,750,000

For disease records that are ordinal variables, non-parametric statistics should be used

Tests for comparing two treatments

Mann-WhitneyWilcoxon Single RankSpearman Rank Correlation CoefficientKolmogrov-SmirnovWald-Wolfowitz Runs

Tests for comparing three (or more) treatments

Krustal-WallisFriedman

Methods for presenting disease records

Treatment Disease severity (%)

Cultivar A 82.5 a

Cultivar B 64.3 b

Cultivar C 50.3 bc

Cultivar D 25.9 c

Cultivar E 4.1 d

Methods for presenting disease records

Time (days)

Dis

ease

sev

erit

y (%

)

=SE

Time (days)

Dis

ease

sev

erit

y (%

)

=LSD

LSD=Least Significant Difference

SE=standard error

There are also problems !

60% ?

Measurement of healthy leaf area

digital analyzers

video camera

measurement of photos

60%

Total leaf area = 125 cm2

healthy leaf area = 98 cm2

diseased leaf area = 27 cm2

Measurement of radiation

Dis

ease

sev

erit

y (%

)Reflected radiation (%)

reflected radiation

observed radiation

Measurement of radiation

Dis

ease

sev

erit

y (%

)

Reflected radiation (%)

radiometer

Concluding remarks

•Disease intensity may be quantified in terms of disease incidence, disease severity or infected area

•The plant organ to be sampled and the disease symptom to be assessed, should be determined according the objective of assessment

•There are several types of scales for disease assessment: descriptive, pictorial, verbal.

•Analysis of the data should be based on the type of variable used - nominal, ordinal or continuous. Calculation of means and conducting ANOVA tests are permitted only for continuous variables.