0

I am trying to conduct a statistical analysis on seed germination data. Effect of 4 different plant extracts on seed germination is tested. Seed germination follows binomial distribution, as it is categorical variable (seed either germinates or not).

Experimental design:

Each of 4 different plant extract treatments have four levels (control and 3 different concentrations) Germination assay is tested in the following way:

  • For each extract concentration 7 petri dishes with 10 seeds is planted (so 70 seeds per treatment concentration). So for each plant extract 70x4=280 seeds in total

Data is collected by counting the proportion of germinated seeds per treatment (example: 4 of 10 seeds germinated) so percentage can be calculated for each treatment.

Basic plot looks like this. In general extracts reduce germination percentage. enter image description here

Data frame is formatted in this way

group germination extract      SD    SE

1 C            64.3 extract 1 1.51  0.571
 2 I            60   extract 1 1.91  0.724
 3 II           48.6 extract 1 1.21  0.459
 4 III          47.1 extract 1 1.11  0.421
 5 C            62.9 extract 2 1.70  0.644
 6 I            71.4 extract 2 1.35  0.508
 7 II           51.4 extract 2 1.07  0.404
 8 III          41.4 extract 2 1.21  0.459
 9 C            58.6 extract 3 1.77  0.670
10 I            60   extract 3 1.63  0.617
11 II           45.7 extract 3 1.99  0.751
12 III          37.1 extract 3 1.11  0.421
13 C            70   extract 4 0.577 0.218
14 I            61.4 extract 4 1.35  0.508
15 II           34.3 extract 4 1.40  0.528
16 III          31.4 extract 4 1.07  0.404

I would like to check whether there is a significant difference between treatment concentrations (control and three concentrations) and whether there is significant difference between plant extracts (to me it seems that it is possible only for certain concentrations, for example is there a difference in germination percentage between third concentration of different extracts).

I know that ANOVA or Kruskal-Wallis cant be used, I have considered Chi Square test (contingency tables for different concentrations), but I am not sure if it can be used for dose - response analysis. My favorite so far is Logistic regression, but I would need some help with R code for that, for determining the right type and parameters.

Maybe something similar to this. I have the original data that represent counts of germinated seed per petri dish per extract per concentration.

What would be the best type of analysis to show the desired differences? Thanks in advance!

EDIT: @John Madden, here is the transformed dataset. I have transformed it to consist of a 1120 rows (each of 4 extracts with three concentration and control, so 4x4x70=1120)

  group germination extract

1 K               1 extract 1
 2 K               1 extract 1
 3 K               1 extract 1
 4 K               1 extract 1
 5 K               1 extract 1
 6 K               1 extract 1
 7 K               1 extract 1
 8 K               0 extract 1
 9 K               0 extract 1
10 K               0 extract 1

and it goes like this K, I, II, III for all 4 extracts, seems that I cant post whole dataset here.

Lestermann
  • 1
  • Welcome to the site! Yes, a logistic regression makes sense (this will allow us to consider the effect of the extract and the concentration simultaneously). To do a logistic regression in R, you'll need to provide your data not in terms of percentages that germinated, but in "long form", i.e. a dataset with 280 rows, each row giving the configuration (so they'rell be lots of duplicates) and with the response vector y consisting of 0's and 1's (if a given seed germinated or not). – John Madden Jun 13 '22 at 15:02
  • also, are you allowed to share what species you're working with? I just like plants :) – John Madden Jun 13 '22 at 15:03
  • @JohnMadden Thank you for your interest. I transformed the dataset. It would be awesome if you could write some R code for logistic regression. Not sure how much can I say, but it is the effect of some moss extracts on common agricultural plant germination. – Lestermann Jun 13 '22 at 17:20

0 Answers0