[Design] Phenotype Importer Columns / Metadata

[laceysanderson]

Current state based on Tripal 3 Raw + Analyzed Phenotypes.

Share (i.e. analyzed)

• Trait Name: The full name of the trait as you would like it to appear on a trait page. This should not be abbreviated (e.g. Days till one open flower).
• Method Name: A short (<4 words) name describing the method. This should uniquely identify the method while being very succinct (e.g. 10% Plot at R1).
• Unit: The unit the trait was measured with. In the case of a scale this column should defined the scale. (e.g. days)
• Germplasm Accession: The stock.uniquename for the germplasm whose phenotype was measured. (e.g. ID:1234)
• Germplasm Name: The stock.name for the germplasm whose phenotype was measured. (e.g. Variety ABC)
• Year: The 4-digit year in which the measurement was taken. (e.g. 2020)
• Location: The full name of the location either using “location name, country” or GPS coordinates (e.g. Saskatoon, Canada)
• Replicate: The number for the replicate the current measurement is in. (e.g. 3)
• Value: The measured phenotypic value. (e.g. 34)
• Data Collector: The name of the person or organization which measured the phenotype.

Collect (i.e. raw)

• Germplasm Name
• Entry Number
• Location
• Plot
• Rep
• Planting Date

[laceysanderson]

Unique identifying the metadata:

1. Germplasm Name
2. Sample Name (i.e. entry no., label on a test tube or pot in the greenhouse)
3. Group (i.e. site/location, assay name)
4. Experimental Unit (i.e. plot, test tube)
5. Replicate
6. Timepoint (e.g. planting date, drone flyover date, assay date)
7. Treatment (e.g. fertilizer, weeding pressure, nitrogen supplementation, temperature)

Field Experiment

1. Variety Name
2. Entry number
3. location
4. plot
5. replicate
6. planting year (not validated though)
7. Examples: nitrogen application, n/a if not needed.

Greenhouse Experiment

1. Germplasm Name
2. Pot label
3. bench name
4. Pot number / label
5. biological replicate
6. planting year / season / quartile
7. same options as in field

Jiayi's experiment

traditional vs. NIR amonio acide composition study

1. Germplasm name (e.g. CDC Asterix AGL)
2. Sample name (e.g. ros 1 sst who)
3. location (e.g. Rosthern, Canada)
4. entry number (number from field)
5. biological replicate from field (i.e. 1-3)
6. 2017 (from the field)
7. assay type (traditional vs. NIR)

LDP 2022 Free SH Group and Soluble Protein Determination (Jocelyn)

Good example of biochemical assay; Did NOT FIT in previous format

1. CDC Crimson (she stored in the sample description)
2. Sample Name (e.g. sample description: 0.50 g Crimson lentil flour in 10.0 mL buffer - heated @ 85ºC)
3. Group (e.g. heated vs. room temperature)
4. Experimental Unit (e.g A1 same as sample name)
5. Replicate (1-3, embedded in sample name)
6. Timepoint (e.g. 2022July13)
7. Treatment (e.g. heated at 85C)

Soluble protein (Still Jocelyn)

1. CDC Crimson (she stored in the notebook)
2. Sample Name (e.g. A1)
3. Group (e.g. dilution factor)
4. Experimental Unit (e.g A1 same as sample name)
5. Replicate (1-3, embedded in sample name)
6. Timepoint (e.g. 2022June8 / Trial No.)
7. Treatment (e.g. heated at 85C, stored in the notebook)

[laceysanderson]

We have discussed this a bit further and come up with the following:

• both phenotypes share + collect will have the exact same format where each trait is a column
• column header for each trait will be a measurement group which uniquely identifies the trait-method-unit combination
• we are still discussing if the measurement group stored as follows:
  • OPTION A: use chado. the measurement group is a cvterm in it's own Genus_group cv. Each group is related to it's trait via a cvterm relationship where the type_id of the relationship indicates this is the trait. The method and unit are also related to the group in the same way via cvterm_relationship
  • OPTION B: this grouping is application specific so instead we create a custom table to store the group. In this approach we will still want to keep the relationship ideal mentioned in option A in order to give us flexibility (i.e. can support more then just trait-method-unit)
• Drop the data collector column from both formats for now. We need to think about the fact that each trait (i.e. column) could be collected by someone else. One suggestion is that we rename this metadata to be the field manager (e.g. Derek).
• Add in the Germplasm Accession Column to both formats but make it optional. This allows us to uniquely identify germplasm better. Note: we know the genus due to configuration of this module and the project associated. However, we do not know the germplasm type so there is still a bit of an assumption going on here. So make sure to have a check that you are pulling out something unique since the key for the stock table is the uniquename - type_id - organism_id and we only provide a way to supply 2 of those.

[ruobinLiu]

Original data collection spreadsheet Downloaded result

We are having multiple values under Plot for a composite mixture of samples in this experiment (Jeremy's B-vitamin quantification in the LR-93 population). He has a mixture of sample from across those three plots present within each analysis. Previously, he did a smaller subset of individual plots, and there weren't significant differences from one plot to another, so this allowed us to analyze the samples in a much more efficient way. If looking for an individual value for any of those plots, then all Plots will return the same vitamin values.