Our approach is guided by:

>Our Objectives (and Value Proposition)

>Our desire to provide value early by taking a step-wise approach to perceived quick-wins

Objective 1 - Be able to compare data across industry partners

First pass and subsequent iterations to populate BASECASE: highlight which assays are most reproducible and robust (possess least variability) as these are most likely to allow early comparison of data across industry partners (avoid need for standardisation in the first instance).
OFFLINE: Use BASECASE table to compile list of meaningful assays for each risk factor. Aim for 5-10 assays.
OFFLINE and ONLINE (PTM): Establish common purpose/description/definition for these assays in context of decision-making vs TPP criteria - aiming to do this PTM first week of July (week 27)
ONLINE (PTM): Discuss and agree properties (e.g. technology used, experimental measurement, variables, identifiers used in the public domain e.g. gene ontology, etc.) to be included and identify key needs e.g. buffer conditions. Aim for 10 max.
Prove that these assays behave the same across partners by testing a selected panel of molecules (not for stakeholder document)
Further comparisons may require a more in-depth analysis of remaining assays, including SOPs (not for stakeholder document)
Attempt to standardise these remaining assays (not for stakeholder document) - remains to be seen if we need to do this

Objective 2 - Identify a minimum set of ‘value’ assays which will provide sufficient information to run a 'standard' campaign.

Define what we mean by a 'standard' campaign; boundaries, milestones and information needed to progress e.g. from Target Product Profile (TPP).
~~Define which information is classed as critical and is associated with developability risk.~~
Identify which assays provide above information and are run as standard.
Prioritise which assays provide most 'value' (bang for buck) information with respect to predictive power (e.g. critical, decision-making, risk data).
Participants would need to identify molecules (range of behaviours), and provide SOPs, guideline metrics and whatever developabilty data exists (from the high-throughput to the detailed) - this requires an agreed upon assay list, and some standardization of assays.
Participants (or contractor, etc) would fill in the missing data (ie likely that bulk of this will include endpoint properties with poorly behaved molecules), using agreed upon assays.
Participants would need to identify how the data will be used, i.e. are we happy with just a qualitative assessment of what correlates, or are we considering something more based on statistical approaches?
Do we want to consider / undertake computational approaches to enable prediction of any of the assays or properties? I would guess that unless we agree to do this, sequence info isn't very critical.
Test, validate and show correlation with endpoint data.
Generate standard assay data for selected (public or internal from 'dead' projects) molecules (index and normalise using standard molecules).
Test if correlations hold for non-standard mAbs.

Objective 3 - Better predict which molecules are likely to succeed based on early biophysical data.

Assess whether or not existing / known assay formats are predictive.

Objective 4 - Predict ideal manufacturing/developability conditions (e.g. formulation) from early biophysical data.

Need SOPs for assays (covered above)
Need sequence data (will be extremely difficult to do this based on structural descriptors alone)
Need to generate more early biophysical data than what is currently done i.e need to run the same assay under different conditions.
Firstly, aim to predict which Abs will exhibit poor behaviours.
Secondly, aim to predict optimal conditions for remaining Abs.

Objective 5 - Define which sequence & 3D structure elements, and biophysics properties contribute to liabilities or undesired effects.

Need to better define structural descriptors (opportunity)

Objective 6 - Predicting endpoint data (e.g. in vivo nonspecific clearance rates) from sequence/structure/structure descriptors.

Establish current internal practices for use of sequence/structure information and structural descriptors to predict developability risk.
Highlight 'gaps' which could be addressed through, more data, better data or better understanding of the data (i.e. correlations)
Combine sequence/structure data with standard 'value' assay data
Seek correlations between sequence/structure and assay data > assay data and endpoint data

Browser not supported