AMReader

About

AMReader is an R package that partners with AMScorer to enable easy quantification of root length colonisation by arbuscular mycorrhizal (AM) fungi, followed by simple data analysis and visualisation.

AMScorer is also distributed with this package and is an excel spreadsheet designed for easy data input alongside microscropy. It has multiple functions intended to enable rapid and accurate data input - ending any dependence on paper, and saving significant data processing time.

AMReader can read AMScorer to instantly perform relevant statistical analyses and data visualisation. Multiple user inputs are built into this function to enable the user to tailor their statistical analyses and graphical outputs. It is raw data to personalised results in a single function.

Installation

Both AMReader and AMScorer can be downloaded from the GitHub repository: EJarrattBarnham/AMReader available at: https://github.com/EJarrattBarnham/AMReader

To download the excel spreadsheet, select “AMScorer.xlsx” from the file list then “Download”.

To download AMReader:

install.packages("devtools")
devtools::install_github("EJarrattBarnham/AMReader")

This requires the R library “devtools”. If having difficulty installing it, try updating the R version you are using to the latest version.

Citation

Please cite the following article if you have used either AMScorer or AMReader in your research. This article contains a description of these tools which is complementary to the user manual below.

Jarratt-Barnham, E., Oldroyd, G. E. D. & Choi, J. Efficiently recording and processing data from arbuscular mycorrhizal colonization assays using AMScorer and AMReader. Front. Plant Sci. 15, (2024). https://doi.org/10.3389/fpls.2024.1405598

In Development

A version of AMScorer which enables concurrent quantification of combinations of fungal structures in the same field of view

Contents

This README provides information on how to use:

1. AMScorer

2. AMReader

AMScorer

AMScorer is designed to be used alongside the R package AMReader, however it can also be used independently.

Introduction

This spreadsheet was developed to aid the collection of data from arbuscular mycorrhizal (AM) root colonisation assays. It is designed to enable rapid data collection, and easy conversion of that data into a graphical representation and statistical outputs.

This spreadsheet contains 26 sheets (A to Z), each of which contains 15 counting tables (Figure 1). Any data input into these is fed into the “AM Results” table, which forms this spreadsheet’s main output. It is also required that metadata detailing the experimental design given in the “Conditions” sheet (which may include information about the genotype, time points, nutrient treatments etc.). Each “Condition” should be given a unique letter in this table.

Figure 1. A counting table

In the first instance, therefore, this spreadsheet enables the counting of up to 15 replicates of 26 conditions. However, it is possible to extend this with the “Blinds” sheet (see section on the “Use of Blinds”) to count a total of 390 plants and a maximum of 78 different conditions (5 replicates per plant in this case), or any other distribution of 390 plants that is desired. It is anticipated that few experiments would exceed this requirement and, if so, multiple copies of this spreadsheet might be complied post-hoc for data analysis.

This spreadsheet also contains a “Test” sheet, which contains 15 counting tables. These results do not form part of any output. If monitoring the colonisation levels of an experiment prior to harvesting, this is a suitable place to record these colonisation data.

Additionally, there is an “Information + Warnings” tab which helps the user reduce the risk of errors in their data recording.

To prevent unintentional changes to the spreadsheet, there is “password” protection. This prevents any cell not intended to be edited from being changed. This can be deactivated by selecting: Home -> Format -> Unprotect sheet. There is no password – just hit enter. Once any changes are made, it is advised you reinstate the password protection.

It is also advised you keep an entirely unaltered copy of the Master Spreadsheet stored and create a duplicate copy for each experiment.

Counting root colonisation

It is our standard approach to record 10 fields of view (at 200 x total magnification) from 10 root pieces per plant. The presence or absence (1 or 0) of Extraradical Hyphae, Hyphopodia, Intraradical Hyphae, Arbuscules, Vesicles and Spores is recorded.

Using these observations, it is possible to quantify the total percentage of colonised root, labelled “Total”, which records the presence of at least one internal structure (Hyphopodia, Intraradical Hyphae, Arbuscules and Vesicles). A count towards “Non” is made if there is no internal structure present. “Total” and “Non” are calculated automatically.

Below is an example of the table for one root piece (Figure 2). This is repeated 10 times per plant.

Figure 2. Data from one root piece

Data input

1. In the “Information + Warnings” tab, fill in the information above the black line with your name and the experiment name.
2. In the “Conditions” tab, insert information about the genotype and growth condition. It is not necessary to give a description for every letter.
3. Go to sheet “A”.
4. From the top-left, open the drop down “named range” menu and select _01 (Figure 3). This will specifically select the boxes to insert the data (see image below for guide). This enables easy data input and is a key feature of this spreadsheet.

Figure 3. Finding the named range

5. Counting AM colonisation.

To input data, few buttons on a keyboard are necessary:

• 1: To indicate the presence of that fungal structure.

• TAB: To move to the next cell in the selection (made in step (4) above).

• SHIFT + TAB: To move back to the previous cell.

• 0: To correct the spreadsheet if you have incorrectly inserted a 1.

On UK/US keyboards, these buttons are all in easy reach with one hand (the left hand is easiest). As most microscopes are set up for use of the right hand to adjust slide positioning, it is advised you use your left hand to input data while you use your right hand to control the microscope. With practice, it is possible to operate this spreadsheet in the corner of your eye. To aid this, cells with “1” in them will go blue.

It is not possible to insert values into the counting tables other than 0 and 1. If you do, a pop-up warning will appear. To close the warning, you can press TAB followed by SPACE or ENTER (or use the cursor to select cancel). This will return the cell to its previous value.

It is possible to accidentally delete a value in a cell. If this happens, the cell will go red to highlight that an error has been made.

6. After completing a slide, move on to the next slide of that “condition”. To do this, select from the top left drop down menu “_02”. This will select the next set of cells. Repeat until all replicates of the given condition are filled in.
7. Once all biological replicates of “A” are complete, move on to sheet “B” for the next condition.

Counting arbuscules and vesicles as intraradical hyphae

Depending on the species and staining quality, it can often be challenging to see intraradical hyphae, though their presence can be assumed by the presence of other fungal structures such as arbuscules and vesicles. Additionally, some researchers consider arbuscules and vesicles as a type of intraradical hypha. This has led to some variability as to whether researchers wish to include a count for intraradical hyphae when they observe arbuscules or vesicles for the purposes of AM colonisation assays. This spreadsheet has therefore been designed to enable easy conversion between the two modes of counting. Under the “Information + Warnings” sheet there is a line “Arbuscules/Vesicles count As Intraradical Hyphae. 1 = YES, 0 = NO”. By submitting “1” into the corresponding box, every field of view containing an arbuscule or vesicle will contribute towards the overall count of intraradical hyphae. By submitting “0”, only the times where a “1” is input in the intraradical hypha column will be counted. As such, if the user ensures they count only what you observe, you can decide to change the intraradical hypha count with a simple switch.

Registering the absence of any fungal structures

In the data analysis stages, slides without any fungal counts are excluded. When counting slides that have no fungus present, this is undesirable. In the event you wish to record a slide that had absolutely no fungal structures, a “No Fungus Override” button (Figure 4) is found on the right-hand side of the counting table. Insert 1 to activate this. Only do this if you want a slide with no fungus to be included in the final results.

Figure 4. The No Fungus Override Button

Warnings

In our experience, few errors are made during counting, and these are easily noticed and readily fixed. However, to ensure that no errors are made, the spreadsheet monitors certain foreseeable errors.

Each counting table monitors whether there are:

1. Any blank cells in the spreadsheet (Counting tables)

Accidentally deleting a value from a cell is an indication of a potential error. If this has occurred, the affected cell will go red, and on the right-hand side of the table, the blank cell warning will activate to highlight which root is affected (Figure 5).

Figure 5. The Blank Cell warning

2. A suspiciously low colonisation count (Counting tables)

It is feasible that, by mistake, a “1” may be inserted into a table that is not intended to be counted. This would lead to the inclusion of these data in the final results. This needs to be avoided. Therefore the “Too Few Counts” warning will activate if:

• Case 1
  • Total colonisation is between 1 and 4.
    AND
  • The sum of extraradical hyphae and spores is less than 4.
• Case 2
  • Total colonisation is less than 4.
    AND
  • The sum of extraradical hyphae and spores is less than 4.
    AND
  • Either extraradical hyphae or spores are greater than 0.

The Too Few Counts warning will therefore catch any accidental inputs of data (Figure 6).

Figure 6. The Too Few Counts warning

3. No Fungus override (Counting tables)

The ability to count a slide as having no fungus may be useful, but would potentially risk inadvertently including “uncolonised” plants in the output for plants that don’t actually exist. If the “No Fungus Override Active” switch is set to 1, it also activates the “no fungus override” warning (Figure 7).

Figure 7. The No Fungus Override Active warning

4. Blind Grading active (Information + Warnings Tab)

If blind grading is active, it is possible that the final results will be altered (see the blind grading section for more information). To ensure this is intended, there is a warning turned on if Blind Grading is active (Figure 8).

Figure 8. The Blind Grading Active warning

Information + Warnings Sheet

If there is any error predicted by the spreadsheet, the “Global Warning” box will go red (Figure 9). The adjacent list of slides will identify the source of the possible error. In the example below, there is an potential error found in slide A1.

Figure 9. The Global Warning

In the case that there are warnings present, it is advised that you check each counting table to determine the cause of the error. If you determine that the error is a false alarm you can deactivate the warning displayed in Figure 9 by informing the spreadsheet that this slide is OK. To do so, insert a “1” into the relevant cell in the “Exemptions” grid (Figure 10).

Figure 10. The Warning Excemptions grid

The Information + Warnings sheet will also let you check which slides have been counted and are contributing to the data (Figure 11).

Figure 11. The Counted Slides grid

Finally, there is also a section to identify which slides have “No Fungus Override” active, so you can easily check these are as you expect (Figure 12).

Figure 12. Slides with No Fungus Override grid

Blinds

It is preferred, and sometimes necessary, for researchers to quantify AM colonisation blind, so as to exclude unconscious bias in their count, or when the genotypes etc. are yet to be determined. To facilitate this, the spreadsheet comes with a “Blinds” sheet (Figure 13). The only column for the user to edit is “Slide (Real) – Specify”. This will give the “true” value of the slide. “Slide (Blind)” records the value of the slide counted as it was named when counting blind. E.g. A researcher may count slide “A1”, then later learn that it belongs to condition “B”. Therefore, put “B” in the “Slide (Real) – Specify” column in the row where “Slide (Blind)” states “A1”. Adding numbers is optional, as the “Condition (Real)” column will remove any numbers from the “Slide (Real) – Specify” column, and this is the data fed into the AM Results table. Acceptable values for the “Slide (Real) – Specify” column are the letters A-Z, AA-AZ and BA-BZ (and any numbers desired).

Figure 13. The Blinds Sheet

If using the “Blinds” sheet, ensure that the “Activate blind scoring” switch (see Figure 13) is set to 1. You can also check the AM Results table to confirm that it is operating correctly.

Blinds – going beyond 26 conditions and 15 replicates

In effect, the “Blinds” sheet informs the output as to which slides are from the same condition. The number of the slide is ultimately not significant. As this allows the sheet to group any set of slides, it is therefore possible to break beyond the 26 different conditions implied by the sheets being labelled A-Z. All 15 replicate tables from each sheet A-Z can be used, and subsequently, with the “Blinds” sheet, they can be grouped independently into up to 78 different conditions (A-Z, AA-AZ, BA-BZ), which can then be named in the “Conditions” tab. This also enables you to count more than 15 replicates of a single condition, as all the counting tables in “B” could be grouped into “A”, by giving them this letter in the “Blinds” tab.

The Output – AM Results

All the data input into this spreadsheet are compiled in the “AM Results” table (Figure 14).

Figure 14. The AM Results table

When a slide has been counted, the “Counted” column become 1 and the line goes blue.

The Output – Conditions

The “Conditions” tab carries the metadata for the “AM Results” tab. Here, you can specify the Condition, Facet_1, Facet_2, Facet_3 and Manual_Colour metadata. An example is given in Figure 15. In this hypothetical experiment there are three different genotypes – WT and mutants 1 and 2. These are listed in the “Condition” column. This experiment tested these genotypes at two different time points (7 weeks and 5 weeks), two different nutrient conditions (Nutrient 1 and Nutrient 2), and two types of inoculum (Inoculum 1 and Inoculum 2) – these are listed in Facet_1, Facet_2 and Facet_3. See “Examples” for illustrations of how this is represented in the final output. Additionally, the “Manual_Colour” tab has been filled in with colour codes that are acceptable inputs in R’s scale_fill_manual() function. These can be hexcodes or any colour listed by the color() function.

Figure 15. The Conditions table

AMReader

The R package AMReader has been designed to:

1. Read the “AM Results” and “Conditions” tabs from the AMScorer and process these data.

2. Perform basic statistical analysis.

3. Produce a graph displaying colonisation percentages.

The goal is to enable the user easy data analysis. Unfortunately, not all user preferences are foreseeable. Consequently, the data tables generated by AMReader are saved to the global environment, allowing the user to conduct their own statistical analyses, and produce their own figures. Additionally, the “Plot” object is also saved to the global environment, allowing the user to fine tune this output, as with any other ggplot object.

There are many parameters in this function, which relate to the three stages listed above. See the “Parameters for” sections below and the AMReader documentation. See “Examples” for examples of how they are employed.

Data Processing

AMReader extracts data from the AM Results and Conditions tab of the excel file. The user is required to provide the “Path” and “File_Name” of AMScorer.

Any slide that has not been counted is excluded from the results.

It may be the user’s preference to include only a subset of the conditions counted in their experiment during statistical analysis and graph production. They can select these using the “Condition_Include” or “Condition_Exclude” parameters.

It may also be the user’s preference to include only a subset of the structures counted in their experiment during statistical analysis and graph production. They can select these using the “Structure_Include” or “Structure_Exclude” parameters.

The user also specifies which Facets to include in the data processing (using “Facet_1”, “Facet_2” and “Facet_3”). In the example from Figure 15, letter code “A” will therefore refer to “WT 7 weeks Nutrient 1 Inoculum 1”. This is important for statistical analysis and graph production.

The key output of these steps is the “Filtered_Dataset”, which is saved to the global environment for the user.

Parameters for Data Processing

The following parameters are important for the Data Processing stages, and may affect the Filtered_Dataset output. Consequently, they also play a role in the Statistical Analysis and Graph Production stages.

• Experiment_Name

  • Choose the Experiment_Name.
  • This is not essential but can play a role at a number of stages:
    • Data import: The default “File_Name” is “AMScorer {Experiment_Name}.xlsx”, where {Experiment_Name} is set by the input of “Experiment_Name”. This gives a quick way of loading the AMScorer without needing to specify the full name of the document.
    • Statistical Analysis output: The default output name is “Statistical Analysis {Experiment_Name}.xlsx”, where {Experiment_Name} is set by the input of “Experiment_Name”. This can allow the output to have a unique tag that refers to the experiment it came from.
    • Graph Production output: The default output name is “Colonisation Percentage Graph {Experiment_Name}.png”, where {Experiment_Name} is set by the input of “Experiment_Name”. This can allow the output to have a unique tag that refers to the experiment it came from.

• Path

  • Set the Path to the folder containing your filled in AMScorer.
  • Any outputs (Statistical Analysis or Graph) will be directed here also.

• File_Name

  • Identify the name of the AMScorer file to import.
  • Defaults to “AMScorer {Experiment_Name}.xlsx”, where {Experiment_Name} is set by the input of “Experiment_Name”.

• Condition_Include

  • Select certain conditions from the data before statistical analysis and graph production.
  • The input is a list of letter codes referring to the conditions to keep in the dataset. E.g. Condition_Include = c(“A”, “C”, “F”) will proceed with only the data from conditions A, C and F.
  • Cannot be used at the same time as Condition_Exclude.

• Condition_Exclude

  • Remove certain conditions from the data before statistical analysis and graph production.
  • The input is a list of letter codes referring to the conditions to remove from the dataset. E.g. Condition_Exclude = c(“B”, “D”, “E”) will exclude any data from conditions B, D and E.
  • Cannot be used at the same time as Condition_Include

• Structure_Include

  • Select certain structures from the data before statistical analysis and graph production.
  • The input is a list of codes referring to the structures to keep in the dataset. E.g. Structure_Include = c(“A”, “V”) will proceed with only the data concerning the abundance of arbuscules and vesicles.
  • Cannot be used at the same time as Structure_Exclude

• Structure_Exclude

  • Remove certain structures from the data before statistical analysis and graph production.
  • The input is a list of codes referring to the structures to remove from the dataset. E.g. Condition_Exclude = c(“EH”, “IH”, “S”) will exclude the data concerning extraradical hyphae, intraradical hyphae and spores.
  • Cannot be used at the same time as Structure_Include

• Facet_1, Facet_2 and Facet_3

  • Inform the function of which facets to take from the Conditions sheet.
  • Options are TRUE or FALSE (default).
  • Please note: If there are two “Condition” values which are identical, as is the case in Figure 15, and the relevant facets are not activated, this will result in all the “WT” conditions being clumped together, ignoring the fact they are not identical. This will mean the the statistics and graph are (likely) meaningless.

Statistical Analysis

By default, AMReader performs no statistical analysis and proceeds straight to producing the graphical output, however AMReader is capable of conducting multiple statistical tests: ANOVA, TukeyHSD, Kruskal-Wallis, Dunn and pairwise Wilcoxon tests. These tests can be used to provide additional information on the graphical output and may also be saved to a “Statistical Analysis.xlsx” output that summarises all these results.

Please note: The statistical tests performed by AMReader are intended to assist the user, but it is not guaranteed that they will be the most appropriate statistical test for the user’s experimental design.

AMReader uses the following R functions for the statistical tests:

• ANOVA – aov()
• TukeyHSD – TukeyHSD()
• Kruskal-Wallis - kruskal.test()
• Dunn Test – DunnTest()
• Pairwise Wilcoxon Tests - pairwise.wilcox.test()

It is recommended the users familiarise themselves with these functions.

Please note: Experimental designs which employ multiple variables (such as in Figure 15) present a challenge to generalised statistical analysis. It is not feasible for the function to take consideration of all possible models for how these variables interact. Instead, AMReader will generate a unique variable name for each condition by combining the “Condition” name and any active “Facet” groups. For example, from Figure 15, condition A becomes “WT 7 weeks Nutrient 1 Inoculum 1”. This is called “Stat_Test_Friendly” and features in the Filtered_Dataset. This allows comparisons between each group of plants that share the same “Condition” and “Facet” values. Analyses of more complex models, and different statistical analyses, must be conducted by the user. Additionally, as part of this processing, any “-” in the user inputs are converted to “_“. This is necessary for the operation of multcompLetters() in the statistical analysis.

Users are able to choose the p-value adjustment methods employed by DunnTest() and pairwise.wilcox.test() using “Stat_Dunn_Padj” and “Stat_Wilcoxon_Padj”.

Users are able to choose whether to conduct one- or two-sided tests using “Stat_Sided”. Unfortunately, the ways in which DunnTest() and pairwise.wilcox.test() process the tests if they are one-sided (either “greater” or “less”) is not intuitive, and operate in different ways. It was decided not to edit these inputs, however, to preserve the “natural” operation of these functions. It is suggested the users familiar themselves with these functions, and examine the outputs of one-sided tests with caution.

These statistical analyses have two main outputs:

• An excel spreadsheet listing the outputs of each statistical test.

  • The first sheet of this spreadsheet will list the input parameters relevant to the statistical tests – the p-value adjustment methods, whether the tests were one- or two-sided, and which conditions and structures were included. This enables the user to know the parameters used to generate the data within the rest of the spreadsheet.
  • ANOVA results, ANOVA diagnostic plots, TukeyHSD results, Kruskal-Wallis results, Dunn test results and pairwise Wilcoxon results are listed in order.
  • Where relevant, cell values < 0.05 are highlighted blue, to enable quick identification of statistical significance.
  • If a reference condition is provided by the user through the “Graph_Reference_Condition” parameter, then any cell containing the text of that reference condition will be highlighted. This will allow easier identification of comparisons of interest in the statistical output.
  • The Statistics_Results data table (within R).
    • This adds additional columns to the Filtered_Dataset stating the statistical significance groups for each structure from each statistical test.
    • If a reference condition is provided by the user through the “Graph_Reference_Condition” parameter, then any condition (Stat_Test_Friendly group) which is statistically different to the given reference condition will also be associated with a “*” in a corresponding column for each structure and statistical test.

The name of the “Statistical Analysis.xlsx” output may be changed from this default using “Stat_File”.

The excel spreadsheet will only be saved if “Stat_Output” = TRUE.

Parameters for Statistical Analysis

The following list describes the options available to the user for personalising the statistical output.

• Stat_Dunn_Padj

  • Choose the p-value adjustment method for the Dunn test.
  • Options are: “none”, “bonferroni”, “holm”, “hommel”, “hochberg”, “BH”, “BY” or “fdr”.
  • The user input is fed into the “method =” parameter of DunnTest().
  • See R DunnTest() documentation for more detail.

• Stat_Wilcoxon_Padj

  • Choose the p-value adjustment method for the pairwise Wilcoxon tests.
  • Options are: “none”, “bonferroni”, “holm”, “hommel”, “hochberg”, “BH”, “BY” or “fdr”.
  • The user input is fed into the “p.adjust.method =” parameter of pairwise.wilcox.test().
  • See R DunnTest() documentation for more detail.

• Stat_Sided

  • Choose whether to perform one or two-sided Dunn and pairwise Wilcoxon tests.
  • Sets of “alternative =” parameters of DunnTest() and pairwise.wilcox.test(). See their documentation for further information.
  • Options are “two.sided”, “less” or “greater”.

• Stat_Output

  • Choose to save the “Statistical Analysis.xlsx” output.
  • Options are TRUE or FALSE (default).
  • The spreadsheet is saved to the folder indicated by “Path”.

• Stat_File

  • Specify the name of the statistical analysis output.
  • Defaults to “Statistical Analysis {Experiment_Name}.xlsx”, where “Experiment_Name” is set by “Experiment_Name”. If Experiment_Name is not given, the output will be “Statistical Analysis.xlsx”.

Graph Production

The majority of parameters from AMReader belong to the production of the graph. These are intended to give the user flexibility as to the aesthetics of the final output.

It is hoped, for the most part, this flexibility will be sufficient for the user. If not, the data used to generate the graph are saved to the global environment for the user to access.

These are:

• Plot_Dataset: A processed version of the Filtered_Dataset.

• Geom_text_information: If statistical analyses have been performed, the outputs selected by the user for graph production are processed to match the Plot_Dataset. The user specifies which statistical results to include using “Graph_Stat_Test” and “Graph_Stat_Display”.

Parameters for Graph Production

The following list describes the options available to the user for personalising the graphical output. See examples for further information on their use.

• Graph_Type

  • The default graph is “Facets” (see examples). This can be changed to “Single”.
  • Options are Graph_Type = “Facets” (defualt) or Graph_Type = “Single”.

• Graph_Sample_Sizes

  • It is possible to include information about the sample size of each condition. E.g. if condition “A” is “WT” and there were six samples, the label can be changed from “WT” to “WT (n = 6)”.
  • Options are Graph_Sample_Sizes = FALSE (default) or Graph_Sample_Sizes = TRUE.

• Graph_Object

  • The default graph is a bar plot (see examples). This can be changed to a box plot.
  • Options are Graph_Object = “Bar” (default) or Graph_Object = “Box”.

• Graph_Datapoints

  • The default option is to display individual data points on the graph. This can be deactivated.
  • Options are Graph_Datapoints = TRUE (default) or Graph_Object = FALSE.

• Graph_Condition_Order

  • It is possible to change the order of the conditions displayed in the graph. This defaults to the alphabetical order of the conditions A-Z etc.
  • It is not necessary to list all conditions, those not specified will preserve their original order after the changes.
  • E.g. Graph_Condition_Order = c(“C”, “E”) will place the condition specified by C and E first, then the remaining conditions A, B, D etc.

• Graph_Facet_1_Order, Graph_Facet_2_Order and Graph_Facet_3_Order

  • If “Facet_1”, “Facet_2” or “Facet_3” are active, it is possible to change the order of these facets, similar to “Graph_Condition_Order”.
  • E.g. From the example of Figure 15, Graph_Facet_1_Order = c(“5 weeks”, “7 weeks”) would ensure that the data for 5 weeks is on the left-hand side of the graph.

• Graph_Stat_Test

  • Choose the statistical test that will provide the information to display on the graph.
  • E.g. If Graph_Stat_Test = “Dunn”, the results of the Dunn Test will be displayed.
  • Options are: “Wilcoxon”, “Tukey”, “Dunn”, or “NULL” (default).
  • If Graph_Stat_Test = NULL, no stats will be displayed on the graph.
  • See also “Graph_Stat_Display”.

• Graph_Stat_Display

  • Choose the representation of the statistical results selected in Graph_Stat_Test.
  • Options are “Letters” (default) or “Reference”.
  • “Letters” will display the compact letter display for statistical comparisons. “Reference” will provide an “*” for statistical significance.
  • Please note: If Graph_Stat_Display = “Reference”, it is necessary to provide information on the condition to compare with - see “Graph_Reference_Condition”
  • Please note: The “*” indicates p < 0.05. There are no further indications of increasing significance.

• Graph_Reference_Condition

  • Choose the reference condition for pairwise comparisons by using the letter code. E.g. Graph_Reference_Condition = “A” will refer to whichever condition is represented by “A”.
  • This is required if Graph_Stat_Display = “Reference”.
  • This also provides information to the statistical analysis, where any pairwise comparison involving the reference condition will be highlighted in the excel output.

• Graph_Manual_Colour

  • To maximise flexibility, the user is able to manually set the colour of every condition in the dataset.
  • This is done through the “Conditions” tab - see Figure 15.
  • Any colour accepted by R - given by the list color() or a hexcode - is valid.
  • Options are TRUE or FALSE (default).
  • Graph_Manual_Colour = TRUE overrides any colour preferences given by “Graph_Colour” and “Graph_Palette” (see below)

• Graph_Colour

  • Other than manually setting the colours, the user can choose to use a colour scheme included in R.
  • Options are “Viridis” (default), “Brewer” and “Grey”, which relate to the functions scale_fill_viridis_d(), scale_fill_brewer() and scale_fill_grey(), respectively.

• Graph_Palette

  • scale_fill_viridis_d() and scale_fill_brewer() have a large number of available colour palettes, which are all accessible to the user.
  • E.g. Graph_Palette = “viridis” will employ the viridis colour palette.
  • See the R documentation for the Viridis and Brewer colour schemes for more information and the available colour schemes.

• Graph_Palette_Begin

  • scale_fill_viridis_d() and scale_fill_grey() map continuous colour scales to discrete bars in the barchart. Users have the option to specify the “range” of colours to choose from each palette - how far to each extreme of the scale to go.
  • “Graph_Palette_Begin” sets the start value of this scale.
  • It takes any value between 0 and 1.
  • If Graph_Palette_Begin > Graph_Palette_End, this will flip the direction of the scale.

• Graph_Palette_End

  • scale_fill_viridis_d() and scale_fill_grey() map continuous colour scales to discrete bars in the barchart. Users have the option to specify the “range” of colours to choose from each palette - how far to each extreme of the scale to go.
  • “Graph_Palette_End” sets the end value of this scale.
  • It takes any value between 0 and 1
  • If Graph_Palette_Begin > Graph_Palette_End, this will flip the direction of the scale.

• Graph_Text_Colour

  • The colour of all text displayed on the graph can be chosen.
  • Any colour accepted by R - given by the list color() or a hexcode - is valid.

• Graph_Background_Colour

  • The colour of the graph background can be chosen.
  • Any colour accepted by R - given by the list color() or a hexcode - is valid.

• Graph_Hline_Colour

  • The colour of the horizontal lines indicating 25, 50, 75 and 100 % can be chosen.
  • Any colour accepted by R - given by the list color() or a hexcode - is valid.

• Graph_Legend

  • Choose to include a figure legend.
  • Options are TRUE (default) or FALSE.

• Graph_Size_Right_Label

  • Modify the size of the text labels on the right hand side of the graph.
  • Only applies if Facet_1, Facet_2 or Facet_3 are active.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Top_Label

  • Modify the size of the text labels on the top of the graph.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Y_Axis

  • Modify the size of the Y-axis label “Percentage Colonisation (%)”.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_X_Axis

  • Modify the size of the X-axis labels giving the conditions.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Legend

  • Modify the size of the figure legend.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Legend_Text

  • Modify the size of the figure legend text.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Percentages

  • Modify the size of the percentage values on the y-axis.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Statistics

  • Modify the size of the statistics information displayed on the graph.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Size_Datapoints

  • Modify the size of the data points displayed on the graph.
  • Acts as a multiplier.
  • Defaults to 1.

• Graph_Output

  • Choose to save the graphical output.
  • Options are TRUE or FALSE (default).
  • The graph is saved to the file indicated by “Path”.

• Graph_Resolution

  • Choose the resolution, in dpi, of the saved graph.
  • Defaults to 1200 dpi for a relatively high quality figure.

• Graph_Width_Adjustment

  • This allows the user to stretch the saved graph’s width in the saved output, allowing more space if necessary.
  • The width is multiplied by the input value (default is 1 - no stretch).
  • The function attempts to adapt the graph width as the size of the graph changes, so the need for this parameter should be relatively limited.

• Graph_Height_Adjustment

  • This allows the user to stretch the saved graph’s height in the saved output, allowing more space if necessary.
  • The width is multiplied by the input value (default is 1 - no stretch).
  • The function attempts to adapt the graph height as the size of the graph changes, so the need for this parameter should be relatively limited.

• Graph_File

  • Specify the name of the graph output.
  • Defaults to “Colonisation Percentage Graph {Experiment_Name}.png”, where “Experiment_Name” is set by “Experiment_Name”. If “Experiment_Name” is not given, the output will be “Colonisation Percentage Graph.png”.
  • Please note: It is necessary to specify the file extension e.g.  “.png”. This allows the user to select the file type of the output (png, pdf, jpeg etc.).

Examples

The following examples use data from AMScorer_Mock_Experiment.xlsx, found in /man/Data. This experiment has the metadata listed in Figure 15. The data in AM Results was generated for use only in this README, they do not refer to any real experiment.

The Simplest Input

The simplest input of AMReader defines the Path and File_Name.

As you will see below, AMReader is relatively verbose in highlighting any parameter that has not been identified by the user, indicating the default values taken by the function. This is intended to help the user see the parameters available to them. For the rest of this README, however, these messages will be silenced.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx"
)
#> [1] "Please provide an experiment name with: Experiment_Name = "
#> [1] "Condition_Include and Condition_Exclude are unassigned. All conditions will appear in the output"
#> [1] "Structure_Include and Structure_Exclude are unassigned. All Structures will appear in the output"
#> [1] "No Stat_Dunn_Padj input, defaulting to Stat_Dunn_Padj = 'none'"
#> [1] "No Stat_Wilcoxon_Padj input, defaulting to Stat_Wilcoxon_Padj = 'none'"
#> [1] "No Stat_Sided input, defaulting to Stat_Sided = 'two.sided'"
#> [1] "No Stat_Output input, defaulting to Stat_Output = FALSE"
#> [1] "No Stat_File input, defaulting to Statistical Analysis {Experiment_Name}.xlsx"
#> [1] "No Graph_Type input, defaulting to Facets"
#> [1] "No Graph_Sample_Sizes input, defaulting to Graph_Sample_Sizes = FALSE"
#> [1] "No Graph_Condition_Order input, conditions will appear alphabetically (given by the condition letter code)"
#> [1] "No Graph_Facet_1_Order input, defaulting to Graph_Facet_1_Order = NULL"
#> [1] "No Graph_Facet_2_Order input, defaulting to Graph_Facet_2_Order = NULL"
#> [1] "No Graph_Facet_3_Order input, defaulting to Graph_Facet_3_Order = NULL"
#> [1] "No Graph_Stat_Test input, defaulting to Graph_Stat_Test = NULL"
#> [1] "No Graph_Stat_Display, defaulting to Graph_Stat_Display = 'Letters'"
#> [1] "No Graph_Reference_Condition input"
#> [1] "No Graph_Manual_Colour input, defaulting to Graph_Manual_Colour = FALSE"
#> [1] "No Graph_Colour input, defaulting to Graph_Colour = 'Viridis'"
#> [1] "No Graph_Palette_Begin input, defaulting to Graph_Palette = 'viridis'"
#> [1] "No Graph_Palette_Begin input, defaulting to Graph_Palette_Begin = 1"
#> [1] "No Graph_Palette_End input, defaulting to Graph_Palette_End = 1"
#> [1] "No Graph_Text_Colour input, defaulting to Graph_Text_Colour = 'grey30'"
#> [1] "No Graph_Background_Colour input, defaulting to Graph_Background_Colour = 'grey95'"
#> [1] "No Graph_Hline_Colour input, defaulting to Graph_Hline_Colour = 'grey30'"
#> [1] "No Graph_Legend input, defaulting to Graph_Legend = TRUE"
#> [1] "No Graph_Size_Right_Label input, defaulting to Graph_Size_Right_Label = 1"
#> [1] "No Graph_Size_Top_Label input, defaulting to Graph_Size_Top_Label = 1"
#> [1] "No Graph_Size_Y_Axis input, defaulting to Graph_Size_Y_Axis = 1"
#> [1] "No Graph_Size_X_Axis input, defaulting to Graph_Size_X_Axis = 1"
#> [1] "No Graph_Size_Legend input, defaulting to Graph_Size_Legend = 1"
#> [1] "No Graph_Size_Legend_Text input, defaulting to Graph_Size_Legend_Text = 1"
#> [1] "No Graph_Size_Percentages input, defaulting to Graph_Size_Percentages = 1"
#> [1] "No Graph_Size_Statistics input, defaulting to Graph_Size_Statistics = 1"
#> [1] "No Graph_Size_Datapoints input, defaulting to Graph_Size_Datapoints = 1"
#> [1] "No Graph_Output input, defaulting to Graph_Output = FALSE"
#> [1] "No Graph_Resolution input, defaulting to Graph_Resolution = '1200'"
#> [1] "No Graph_Width_Adjustment input, defaulting to Graph_Width_Adjustment = 1"
#> [1] "No Graph_Height_Adjustment input, defaulting to Graph_Height_Adjustment = 1"
#> [1] "No Graph_File input, defaulting to Colonisation Percentage Graph {Experiment_Name}.png"
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx"): Facet_1 = FALSE, but more than one unique
#> variable has been found for Facet_1. This would suggest that Facet_1 should be
#> TRUE. Please check this and amend if necessary.
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx"): Facet_2 = FALSE, but more than one unique
#> variable has been found for Facet_2. This would suggest that Facet_2 should be
#> TRUE. Please check this and amend if necessary.
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx"): Facet_3 = FALSE, but more than one unique
#> variable has been found for Facet_3. This would suggest that Facet_3 should be
#> TRUE. Please check this and amend if necessary.

Alternatively, we can access the file using the Experiment_Name parameter.

library(AMReader)
AMReader(
  Path = "man/Data",
  Experiment_Name = "Mock Experiment"
)

Additionally, AMReader is designed to pre-empt errors, and guide the user to correct them.

In the above example, 3 warnings are activated. This is because the experimental design contained three different Facets (Figure 15).

The Use of Facets

In the above graph, all the data have been grouped together under “WT”, “Mutant 1” and “Mutant 2” because none of the Facets have been activated.

AMReader detects this as likely unintended, and guides the user correct this.

For example, from the above: Facet_1 = FALSE, but more than one unique variable has been found for Facet_1 in Conditions. This would suggest that Facet_1 should be TRUE. Please check this and amend if necessary.

Correcting this:

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = TRUE,
  Facet_2 = TRUE,
  Facet_3 = TRUE
)

Now the output is divided into each individual “Faceted” combination from the metadata. This represents all the data present in the results.

Selecting Structures and Conditions

Evidently, these results are not so easy to interpret, since too many figures are displayed. We can therefore choose certain Structures and Conditions to display using Structure_Include or _Exclude and Condition_Include or _Exclude.

For example, we can choose to display only “Total”, “Arbuscules” and “Vesicles” from structures, and only data from “7 weeks”. From Figure 15, we can see that the letter codes referring to “5 weeks” are letters M to X, and we can choose to exclude these.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = TRUE,
  Facet_2 = TRUE,
  Facet_3 = TRUE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Exclude = c("M", "N", "O", "P", "Q", "R", 
                        "S", "T", "U", "V", "W", "X")
)
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx", : Facet_1 = TRUE, but no more than one unique
#> variable has been found for Facet_1. This would suggest that Facet_1 should be
#> FALSE. Please check this and amend if necessary.

This makes our data much easier to visualise. Now we can see that the Inoculum facet makes a limited impact on colonisation, while there is a clear effect of the two Nutrient facets.

Additionally, we no longer need to apply the Facet_1 condition, as we are only visualising the “7 weeks” data. AMReader has detected this and given a corresponding warning in the above output.

We may now choose to display only the data for 7 weeks and Inoculum 1. Instead of listing all conditions to “exclude”, we can instead choose to “include” these conditions, making the list of letters shorter.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I")
)

Using the above parameters, therefore, allow us to flexibly choose which data to present in our final figure.

Statistics

Generating the Statistical Analysis.xlsx output

Now, we may wish to display some statistical information on the graph.

To perform these analyses, we would first wish to generate the Statistical Analysis.xlsx output, and investigate the statistical results.

We will, in this instance, perform two-sided tests, without any p-value adjustment. This is the default, but we will specify the parameters in this case.

To generate a unique output name, we will use the “Stat_File” parameter.

To save the output, we will set Stat_Output = TRUE.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = TRUE
)

Please see the excel output in /man/Data.

The statistical analysis is only conducted for Structures and Conditions set in the above function. If you wish to have comprehensive set of results, the user might wish to conduct the statistical analysis without selecting any specific structures and conditions, save the results, set Stat_Output to FALSE to prevent these results being overwritten (or change Stat_File), and then focus on generating the graph of interest.

Please note: We will not discuss the statistical details here. The output generated by AMReader is intended to be a useful tool, but it is the user’s responsibility to confirm that the statistical tests are appropriate.

Displaying statistics on your graph

Having conducted these analyses, we may, for example, wish to display the results of the Dunn tests on our graph output.

We can do this using the Graph_Stat_Test parameter.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn"
)

Now the graph shows the compact letter display from the Dunn test output.

We may also wish to visualise statistical differences compared to a given reference condition. The default is Graph_Stat_Display = “Letters”. This can be changed to Graph_Stat_Display = “Reference”.

If using Graph_Stat_Display = “Reference”, it is necessary to indicate which letter code is the comparison group. This is done with Graph_Reference_Condition. Let us, in this case, compare to “A” - WT Nutrient 1.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Reference",
  Graph_Reference_Condition = "A"
)
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx", : Graph_Stat_Display = 'Reference' may be
#> misleading if Facet_1/2/3 are active. See manual for further information.

Now we can see which comparisons to WT in Nutrient 1 are significant (p < 0.05).

Please note: The Graph_Stat_Display = “Reference” option is naive to the use of facets. It cannot compute comparisons to the “WT” within each facet. Hence, we can see that Mutant_1 in Nutrient 2 displays a significant difference compared to the WT in Nutrient 1. When Facets are active, therefore, Graph_Stat_Display = “Reference” is often misleading, and “Letters” should be used instead. A warning has been activated in the example above because of this.

Please note: Graph_Stat_Display = “Reference” will not display any additional asterisks for increasing levels of significance.

Adding sample size information

In addition to the statistics, we may wish to convey the sample size of each condition.

We can do this with Graph_Sample_Sizes = TRUE.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE
)

Graph Aesthetics

We may also wish to with adjust the appearance of the graph.

Figure Legend

We may decide, for example, that we do not need to display the figure legend.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE
)

Colour Schemes

We may also wish to change the colour scheme.

With Graph_Colour, we can choose one of “Viridis”, “Brewer” and “Grey” colour schemes, getting access to their multiple colour palettes. We can choose the palette with Graph_Palette.

For example, we can choose the “blues” palette from Brewer.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Brewer",
  Graph_Palette = "Blues"
)

We might also wish to provide a personal colour scheme.

We can do so by setting Graph_Manual_Colour = TRUE. This will pull metadata from the conditions tab (see Figure 15). This will override any inputs from Graph_Colour and Graph_Palette.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Brewer",
  Graph_Palette = "Blues",
  Graph_Manual_Colour = TRUE
)
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx", : Graph_Manual_Colour = TRUE overrides any
#> Graph_Colour input

Going forward, we will return these bar colour options to the defaults: Graph_Colour = “Viridis”, Graph_Palette = “viridis”, and deactivate the Graph_Manual_Colour option.

The colour schemes “Viridis” and “Grey” work by choosing parts of a continuous colour scale which are then mapped to discrete values to colour the bars. This scale can be represented by the range 0-1.

We can choose a particular part of this scale to colour our graph using Graph_Palette_Begin and Graph_Palette_End.

For example, we may wish to avoid the extreme ends of the viridis palette by setting these to 0.2 and 0.8.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.2,
  Graph_Palette_End = 0.8
)

These parameters are also useful if we wish to change the direction of the colour scheme.

For example, if we switch Graph_Palette_End to 0.2, and Graph_Palette_Begin = 0.8, we flip the colour scheme.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2
)

We may now also wish to change the text colour. Perhaps we would prefer black to the default “grey30” colour.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black"
)

Condition and Facet Order

We also have the opportunity to change the condition order using Graph_Condition_Order and the colour codes of each condition (see Figure 15).

For instance:

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black",
  Graph_Condition_Order = c("C", "B", "A", "F", "E", "D")
)

This has overriden the default alphabetical ordering of conditions, though the facets are still preserved.

It is not necessary to specify all letters with Graph_Condition_Order. Any letters not mentioned will default back to alphabetical order.

For example:

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black",
  Graph_Condition_Order = c("C", "F")
)

This brings Mutant_2 (C and F) to the left-hand side of the graph, but WT (A and D) then appear before Mutant_1 (B and E).

In addition to the order of conditions within each facet, we can change the order of facets themselves using Graph_Facet_1_Order, Graph_Facet_2_Order and Graph_Facet_3_Order.

Since Nutrient 1 and Nutrient 2 are from Facet_2, we can switch these around as follows:

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black",
  Graph_Condition_Order = c("C", "F"),
  Graph_Facet_2_Order = c("Nutrient 2", "Nutrient 1")
)

These examples above illustrate significant flexibility with the style of the graphical output we produce. However, in each of the above cases, each “Structure” has been displayed within their own separate facet. We may, instead, wish to display the all the structures for one condition side-by-side.

Graph_Type: Single

We can do this by setting Graph_Type = “Single”. The default we have been using before was “Facets”.

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black",
  Graph_Condition_Order = c("C", "F"),
  Graph_Facet_2_Order = c("Nutrient 2", "Nutrient 1"),
  Graph_Type = "Single"
)
#> Warning in AMReader(Path = "man/Data", File_Name =
#> "AMScorer_Mock_Experiment.xlsx", : Statistics cannot be displayed if Graph_Type
#> == 'Single'. Choose Graph_Type = 'Facets' to see statistics

Regrettably, statistical information cannot be mapped properly with this graph type. You may observe the corresponding warning in the output above.

All other parameter choices, however, will be preserved.

Text sizes

As we change the data and format of the graph, AMReader will attempt to change and adapt the size of any text and the data points displayed to give a good output. It cannot do this perfectly, however. Consequently, the size of all text, and the data points on the graph, can be altered by the user manually.

They can do this using a series of “Graph_Size_” parameters.

• Graph_Size_Right_Label
• Graph_Size_Top_Label
• Graph_Size_Y_Axis
• Graph_Size_X_Axis
• Graph_Size_Legend
• Graph_Size_Legend_Text
• Graph_Size_Percentages
• Graph_Size_Statistics
• Graph_Size_Datapoints

These parameters act as multipliers. A value of 2 doubles the size of the text, a value of 4 quadruples the size of the text.

For the sake of brevity, we will not demonstrate these parameters here. You may look up their functions in more detail in the “Parameters for Graph Production” section.

Saving the Graph

Finally, we may wish to save the graphical output to our computer.

To do this we would:

• Set Graph_Output = TRUE.
• Choose the resolution of the output (in dpi) with Graph_Resolution.
• Name the file output with Graph_File
  • It is necessary to specify the file extension in this name, this allows the user to choose the file format that is saved (e.g. .png, .pdf etc.)

For example:

library(AMReader)
AMReader(
  Path = "man/Data",
  File_Name = "AMScorer_Mock_Experiment.xlsx",
  Facet_1 = FALSE,
  Facet_2 = TRUE,
  Facet_3 = FALSE,
  Structure_Include = c("Total", "A", "V"),
  Condition_Include = c("A", "B", "C", "G", "H", "I"),
  Stat_Sided = "two.sided",
  Stat_Dunn_Padj = "none",
  Stat_Wilcoxon_Padj = "none",
  Stat_File = "Statistical_Analysis_Mock_Experiment",
  Stat_Output = FALSE,
  Graph_Stat_Test = "Dunn",
  Graph_Stat_Display = "Letters",
  Graph_Reference_Condition = "A",
  Graph_Sample_Sizes = TRUE,
  Graph_Legend = FALSE,
  Graph_Colour = "Viridis",
  Graph_Palette = "viridis",
  Graph_Manual_Colour = FALSE,
  Graph_Palette_Begin = 0.8,
  Graph_Palette_End = 0.2,
  Graph_Text_Colour = "black",
  Graph_Condition_Order = c("C", "F"),
  Graph_Facet_2_Order = c("Nutrient 2", "Nutrient 1"),
  Graph_Type = "Single",
  Graph_Output = TRUE,
  Graph_File = "Colonisation_Mock_Experiment.png",
  Graph_Resolution = 1200
)

The last two parameters we have yet to cover are Graph_Width_Adjustment and Graph_Height_Adjustment. These will change the plotting size of the saved image. This allows the user to stretch the output’s width and height. These parameters act as multipliers. A value of 2 doubles the size of the dimension, a value of 4 quadruples the size of the dimension.

Happy Counting!

Edwin Jarratt-Barnham.