Chapter 3 Worked examples

In this chapter, short examples are given on how to apply the ASPAR_KR metadata standardisation programme. A download link to the standardised excel sheet is given at the end of each section.

3.1 Air Samples according to Kortenbosch et al. (2022).

“Following the example of Peter, we see what it takes to fill in the meta data for a limited air sample study”.

Kortenbosch et al. (2022) reports a method of cheaply taking air samples of A. fumigatus. In this technique, Delta traps are assigned to observation units and left for a certain number of days. Then the Delta traps are transferred to a flamingo agar (Zhang et al. (2021)) plate with azole, and without and the colony forming units are counted.

Peter is planning to determine the azole resistance of A. fumigatus in Arnhem vs Nijmegen. Per town, he hangs 4 air sample units, and he plans to plate them on flamingo agar and flamingo agar with itraconazole (4 mg/L).

To record these measurements using ASPAR_KR, take the following steps:

3.1.1 Set up the metadata sheet

  1. Go to https://aspar-kr.bioinformatics.nl/template. And fill in step 1 according to Figure 3.1.
    1. Fill in Investigation information:
      • Identifier: Arnhem_vs_Nijmegen;
      • Title: Azole resistance in A fumigatus in Arnhem compared to Nijmegen.
      • Description: The Itraconazole resistance fraction was studied using the airsampling method of Hylke et al. The Fractions did not differ much. Four airsampling units were placed in each town.
    2. Fill in person information (click add when finished).
      • First name: Peter.
      • Last name: Doe.
      • email: .
      • Organisation: Example company.
      • Role: Being a good example.
      • ORCHID: <left empty>.
      • Department: Example department
Filled in investigation information

Figure 3.1: Filled in investigation information

  1. Add Study information (Click: Study information).
    Since the design is simple, we can copy the information from investigation.

  2. Add the standard observational unit template.
    Click: Observation unit information > Search a package > default.

  3. Fill in the experimental metadata.
    Since we need to keep track of Delta traps, and their corresponding culture we use airstrip samples, and CFU count samples.

    1. Add the AirStrip sample package:
      Click: Sample Information > search a package > Scroll down and add AirStrip > Add template.
    2. Add the CFU sample package:
      Click: Sample Information > search a package > Scroll down and add CFUCountCulture > Add template.
    3. Adjust your optional metadata requirements.

  4. No assay information is needed since Peter did not plan to do genotyping or another test.

  5. Download the excel sheet (Generate workbook). You can download the resulting sheet here, to circumvent steps 1-5.

3.1.2 Fill in the data

3.1.3 The common metadata tables

As mentioned in Section 2.1, between each class, objects are linked by identifiers. In the Arnhem_vs_Nijmegen.xlsx sheet things are no different, you’ll find the following 5 tables in this sheet:

  1. Table 3.1: The investigation. Links to Study via investigation identifier.

  2. Table 3.2: The study. Links to observational unit via study identifier.

  3. Table 3.3: The observational unit. Links to Sample via observational unit identifier.

These three tables will always be present. The actual research data will is always part of the sample and assay tables.

Table 3.1: Investigation table. Investigation has the investigation identifier column.
investigation identifier investigation title investigation description firstname lastname email address orcid organization department
Arnhem_vs_Nijmegen Azole resistance in A fumigatus in Arnhem compared to Nijmegen. The Itraconazole resistance fraction was studied using the airsampling method of Hylke et al. The Fractions did not differ much. Four airsampling units were placed in each town. Peter Doe org s
Table 3.2: The Study table.
study identifier study description study title investigation identifier
Arnhem-vs-nijmegen 4 samples in each town were made using hylkes method Azole resistance in A fumigatus in Arnhem compared to Nijmegen. Arnhem_vs_Nijmegen
Table 3.3: Observational unit table.
observation unit identifier observation unit name observation unit description study identifier
arnhem Arnhem Measurements in Arnhem Arnhem-vs-nijmegen
nijmegen Nijmegen Measurements in Nijmegen Arnhem-vs-nijmegen

3.1.4 Experimental data tables

3.1.4.1 Air samples

To fill in these tables Peter hangs the Air sample strips up for 4 weeks: between the 5th of November and the 26th of November. He records the minimal information for the air sample strips:

  1. start end end date.

  2. Collection location.

    • Lat
    • Long

For taxonomy ID he picks 32644: unidentified, since Peter does not know which species he will collect on his Delta Trap. For the culture made using the Delta Trap, he selects for A. fumgiatus, so 746128

He also collects the recommended and optional data:

  • Altitude
  • Environmental context:
    • Broad scale.
      Overall environment where the sample is found? What has the biggest influence on the environment of the sample?
    • Local scale.
      Things in the vicinity of the sample that may influence the outcome of the experiment.
    • Environmental medium.
      Things that directly surrounded your sample prior to sampling.

In Arnhem & Nijmegen he hangs some samples in trees of the the park, others at the train station at traffic poles. Since he wants to record these terms as part of the environmental contex, he looks up the terms on ontobee/envo (Figure 3.2). He finds that:

Lookup of a machine readable term [on Ontobee](https://ontobee.org/search?ontology=ENVO\&keywords=station\&submit=Search+terms).

Figure 3.2: Lookup of a machine readable term on Ontobee.

Using this information, he fills in the Delta Trap excel sheet.

3.1.4.2 Culture Samples

After filling in the air sample sheet, he fills in the information for the two layer culture:

As a selection medium, he adds itraconazole to a concentration of 4 g/L (\(x\) molar)1, which has CHEMBL1835949 as the standard identifier.

3.1.5 Convert the sheet to ttl format

  1. Go to https://aspar-kr.bioinformatics.nl/validate.
  2. Upload the finalised sheet.
  3. Click “DOWNLOAD RDF”.

3.1.6 Analysis of the FAIR data with R

To analyse the data we can use the rdflib package in R together with the tidyverse.

base::dir.create("hylke_air_method_example")
destfile="./hylke_air_method_example/data.ttl" 
 fileURL <-
 "https://git.wur.nl/aspar_kr/aspar/-/raw/main/example_files/Arnhem_vs_Nijmegen.ttl?ref_type=heads"   
 if (!base::file.exists(destfile)) {
    utils::download.file(fileURL ,destfile,method="w")
 }
 
rdf <- rdflib::rdf_parse("hylke_air_method_example/data.ttl",
                         format = "turtle")
rdf
## Total of 416 triples, stored in hashes
## -------------------------------
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_nijmegenAirSamples/sam_CultureNijmegenStation2> <http://fairbydesign.nl/ontology/biosafety_level> "2"^^<http://www.w3.org/2001/XMLSchema#integer> .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_nijmegenAirSamples/sam_CultureNijmegenStation2> <http://schema.org/description> "A two layer culture made from the delta trap taken from the station square in Nijmegen" .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen> <http://schema.org/identifier> "arnhemVsNijmegen" .
## <http://fairbydesign.nl/ontology/selection_medium> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/1999/02/22-rdf-syntax-ns#Property> .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_nijmegenAirSamples/sam_CultureNijmegenPark1> <http://fairbydesign.nl/ontology/antibiotics> "CHEMBL1835949" .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_arnhemAirSamples/sam_CultureArnhemStation2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://jermontology.org/ontology/JERMOntology#Sample> .
## <http://fairbydesign.nl/ontology/biosafety_level> <http://schema.org/valueRequired> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_nijmegenAirSamples/sam_CultureNijmegenStation2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://jermontology.org/ontology/JERMOntology#Sample> .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_nijmegenAirSamples/sam_CultureNijmegenStation2> <http://fairbydesign.nl/ontology/medium_type> "flamingo medium" .
## <http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/stu_arnhemVsNijmegen/obs_arnhemAirSamples/sam_arnhem2> <http://fairbydesign.nl/ontology/packageName> "DeltaTrap" .
## 
## ... with 406 more triples

After reading the RDF object we can perform simple SPARQL queries to get tibble data.

sparql_query <- 
  ' # Select the first 5 of everything...
  SELECT * WHERE {?s ?p ?o} LIMIT 5
  '
result <- rdflib::rdf_query(rdf, sparql_query)
result
## # A tibble: 5 × 3
##   s                                                                  p     o    
##   <chr>                                                              <chr> <chr>
## 1 http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/st… http… 2    
## 2 http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/st… http… A tw…
## 3 http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/st… http… arnh…
## 4 http://fairbydesign.nl/ontology/selection_medium                   http… http…
## 5 http://fairbydesign.nl/ontology/inv_arnhemVsNijmegenComparison/st… http… CHEM…

Now, to plot an overview of Peter’s experiment we can use:

sparql_query <- 
  
  ' # Select the first 5 of everything...
  prefix ppeo:     <http://purl.org/ppeo/PPEO.owl#> 
  prefix jerm:     <http://jermontology.org/ontology/JERMOntology#> 
  prefix fair:     <http://fairbydesign.nl/ontology/> 
  prefix rdfs:     <http://www.w3.org/2000/01/rdf-schema#> 
  prefix schema:   <http://schema.org/>
  SELECT ?observation_label ?sample_label (COUNT(?sample_label) as ?n) ?total_cfu ?selection_cfu WHERE {
    # Get the samples of interest
    ?observation_unit a ppeo:observation_unit .
    ?observation_unit jerm:hasPart ?parts .
    ?parts a jerm:Sample .
    ?parts fair:packageName "TwoLayerCulture" .
    ?observation_unit schema:name ?observation_label .
    ?parts schema:name ?sample_label .
    
    # Experimental data
    ?parts fair:total_cfu ?total_cfu .
    ?parts fair:selection_cfu ?selection_cfu .
  } GROUP BY ?observation_unit
  '
result <- rdflib::rdf_query(rdf, sparql_query)
result
## # A tibble: 2 × 5
##   observation_label    sample_label                    n total_cfu selection_cfu
##   <chr>                <chr>                       <dbl>     <dbl>         <dbl>
## 1 The city of Arnhem   Arnhem Centraal culture         4        66            28
## 2 The city of Nijmegen Nijmegen stationsplein cul…     4        57            21

The result table, we can modify and plot using the tidyverse.

result |> 
  dplyr::mutate(resistance_fraction = selection_cfu / total_cfu) |> 
  ggplot2::ggplot(ggplot2::aes(x = observation_label, y = resistance_fraction)) +
  ggplot2::geom_bar(stat="identity")

References

Kortenbosch, Hylke H., Fabienne Van Leuven, Bas J. Zwaan, and Eveline Snelders. 2022. “Catching More Air: An Effective and Simple-to-Use Air Sampling Approach to Assess Aerial Resistance Fractions in Aspergillus Fumigatus.” Preprint. Microbiology. https://doi.org/10.1101/2022.11.03.515058.
Zhang, Jianhua, Alfons J. M. Debets, Paul E. Verweij, and Sijmen E. Schoustra. 2021. “Selective Flamingo Medium for the Isolation of Aspergillus Fumigatus.” Microorganisms 9 (6, 6): 1155. https://doi.org/10.3390/microorganisms9061155.

  1. Molecular weight of itraconazole is 705.6 g/mol, he added 4 g to 1 liter of flamingo medium. Therefore ((1 * mole) / (705.6 * gram)) * (4 * gram) = approx. 5.6689342 mmol.↩︎