Menu
▾
▴
#809 assay harmonization - device
46 of 175 assays have defined devices used in the assays.
List of assays with defined devices (shown in the ‘device’ worksheet):
https://docs.google.com/spreadsheets/d/1wg2Gg2OvnkX84lVeqRO0ie-tSvtEe4Pq0rtnKsveas/edit?usp=sharing
They almost primarily use has_participant with only one assay, ‘mass spectrometry assay’, using has_specified_input. It should also be represented in same way as the other devices are, using has_participant.
IEDB assay defined device used in an assay as:
(has_specified_input some X) and (realizes some (function and ('inheres in' some X)))
No specified fuction has been used in the logical axiom. It may a way to distinguish device input from other kinds of input materials.
We may consider to create specific relations for different input of assay to simplify the logica axioms, such as 'use device' for device, 'use reagent' for reagent input.
Some assays have multiple similar function devices specified in their axioms, such as ‘chain termination sequencing’ (http://purl.obolibrary.org/obo/OBI_0000695) with the device ‘ABI 377 automated sequencer’ or ‘Li-Cor 4300 DNA Analysis System’. This may cause a maintenance issue in the future. It can be considered to make an organization device class, such as chain termination sequencer.
We should try to keep the logical definitions of device, evaluant, reagent etc. as symmetric as possible. For evaluant we had as the definition template:
(has_specified_input some X and 'has role' 'evaluant role') and (realizes some ('evaluant role' and ('role of' some X)))
The equivalent for device would be:
(has_specified_input some X and 'has function' some Y) and (realizes some ('function Y' and ('function of' some X)))
This differs from the above by wanting to specificy both the function and the device. It needs to be clarified how this is best implemented in a spreadsheet, as multiple variables are required. One possibility is to extend the ROBOT syntax. Another is to have the device + function as an OWL expression in the cell of the spreadsheet.
As an example, an analytical chromatography assay, which utilizes a chromatography column to separate in input material into components should be modeled as:
(has_specified_input some chromatography column and 'has function' some material separation function) and (realizes some ('material separation function' and ('function of' some chromatography column)))
The need for the two part statement (the device being the input and the function being realized) is due to limitations in OWL which do not allow us to state that the same device participating in the assay is bearing the function that is realized in the assay. The annoyance of this is mitigated by the fact that we can use spreadsheets to make these statements.
Need feedback including from James O regarding implementation in ROBOT.
Call 9/12: James confirmed that while the 'double variable' specification is not currently possible in ROBOT, extending it should be feasible. Decision to implement as specified above.