Discovery campaign analysis
Ten files of assay data from a discovery project have been delivered by a CRO. The agent is prompted to integrate the data sets and to deduce the thresholds used to promote clones down the discovery funnel.
Integrate the campaign’s exports
A campaign arrives as the assays were run: primary ELISA, hit sequencing, Fab expression, SPR off-rate, IgG conversion and production, flow cytometry, IC50 blockade, developability, lead nomination. Each file holds the clones that reached that stage, in its own layout. The agent joins them on clone identity into one workbook table, so every candidate keeps one row and the assays it never reached stay empty rather than dropping it from the analysis.
Read the promotion funnel
Which clones were promoted is recorded only by which of them appear in the next export. Comparing the sets stage by stage gives the funnel, and comparing the readouts of the clones that advanced with those that stopped gives the thresholds that separated them. Those are the gates the campaign ran on, whether or not anyone wrote them down. The chart shows every clone against every assay axis, with the survivors drawn through it.
Keep the evidence with the conclusion
A threshold read from the data is a proposal, not a documented criterion, and the workbook says which it is. Each export is recorded with what was taken from it, a gate is kept beside the clones it does and does not explain, and where the data cannot settle a question, such as two assays that would exclude the same clone or a cutoff that several values fit, that is stated rather than resolved.
Customize the analysis
The same workbook takes a campaign of any shape: different assays, more stages, several targets at once, or an organization’s own promotion rules in place of the inferred ones.