1 of 1 people found this helpful
Looks like you have a problem here:
-activity in cell line A (mean, standard error of mean and n number- 3 fields)
- activity in cell line B, cell line C etc.
That reads like you need another table with one record for each cell line that is in turn linked to your Compounds table.
Need to get the data-model right before you can figure out aggregate calculations based on that same data.
Hi Phil, yes, sorry, I do have a cell lines table which has a unique field for the cell line. How should this be related to the other tables though?
It would only be related to compounds from what I have read thus far.
What is not clear to me is the "paper relationship" between experiments and cell lines. Your current data model assumes that any experiment needs to be equally linked to all of a compound's cell lines and that any results would apply equally to each line.
That seems unlikely.
I'm wondering if you might not have a different Assay_Entry for each cell line for a given compound. If so, that cell line data might well become additional fields in the Assay_entry table.
this is where my understanding of filemaker relationships is lacking.
Each experiment uses one of five different cell lines, and on each experiment I can test up to a maximum of three compounds.
As it stands, my assay_sample join table should give a set of records which have cell line data (copied from the experiment table), the compound number(fk) and the experiment number(fk). I then want to divide that set up according to cell line and perform calculations on the separate lists.
OK, but does one Assay_Sample record record information specific to a single cell line or multiple cell lines?
And I need to walk back that speculation about putting the cell line data into that join table as fields. That clearly isn't a good idea here as I'm forgetting that multiple experiments can be performed on a given cell line.
each assay_sample can only be associated with one cell line since each assay_sample is associated with a single assay_experiment and each experiment is associated with one cell line.
OK, my assumptions are confirmed. That suggests this data model for part of your relationship graph:
Compound;;__pkCompoundID = Cell_Lines::_fkCompoundID
Cell_Lines::__pkCell_LineID = Assay_Entry::_fkCell_LineID
what do the the ;;, :: and pk parts mean?
I have some typos there. Every semi-colon should be a colon.
In FileMaker, standard data references are in this format: TableOccurrenceName::FieldName. A Table Occurrence is a "box" in your relationship graph. It represents a table in your solution, but you can have many Table Occurrences in your graph that represent the same table.
pk stands for "primary key". fk stands for Foreignkey. These are my preferred naming conventions. I put __ in front of pk and use it in no other field name so that it alphabetically sorts to the top of lists that are found in many places in FileMaker when doing design work. A single _ puts all fk fields just below the pk field. This is just my preference, there are many different naming conventions that developers use.
makes sense! let me play around with it a while and then hopefully I can figure out how to do the subset record thing.
many thanks for your help!