Engineer Proteins

The Engineering Pad (ePad) enables users to design new Fvs or Building Blocks from existing Fvs/BBs in BioGlyph, conditioned on in silico properties calculated within the software. In BioGlyph, an Engineering Submission is an iterative process of 1. Selection of Fvs/BBs based on predicted in silico properties and 2. Design of new Fvs/BBs, which become the basis for selection in the next Engineering Cycle.

Engineering Cycles

In BioGlyph, Engineering is an iterative cycle of Filtering & Selecting and Engineering.

Using predicted in silico molecular properties (and experimental properties if available), users can filter entities down to those matching a desired property profile. By applying mutations in a property-guided manner to the set of selected entities, users are able to engineer new entities which can be filtered and selected to get the next round of selections.

This process of Filtering & Selecting followed by engineering is referred to as a Cycle (Cycle counts start at 1). Generations indicate how far from the initial set of parent entities an entity is. In the first cycle, entities which already exist in the request (Generation 0) are filtered down to a Selection using the Engineering Decision Matrix. In the ePad, users apply mutations to the Selection set of parent entities to generate variants. These variants are the N+1 Generation where N is the Generation number of the variant’s respective parent. Both the variants and parents are pushed into the next Cycle, where the user can begin a new iteration of Filtering & Selecting and Engineering. Once a user completes a Cycle, the selection for that cycle is locked. However, users can “fast forward,” or pull entities from past cycles into the current cycle, by changing the cycle filter to point to previous cycles.

Create New Submissions

Engineering Submissions are currently done on a Target and Type basis, meaning that in a single submission, you may only engineer Fvs or BBs of the same type (e.g. Fv or VHH) and target. On the first page of the Engineer Proteins task, the user must select the Target, Type, and Objective (Developability, Affinity Modulation, Humanization, or Other). A description is optional. The list of enabled types in the Type dropdown are based on their presence in formats saved to the current Request. After selection, the user hits the “Start Engineering” button to proceed to the initial selection step (Decision Matrix).

The user can view previous Engineering Submissions by clicking on the “Submissions” button in the top right corner. From the Submissions page, one can access the previous submissions by type, objective, or status and jump to the latest Engineering Cycle. The status shown indicates whether an Engineering Submission has been finalized or if it is still ongoing.

Engineer Decision Matrix - Select Fvs/BBs

Engineering Submissions are currently done on a Target and Type basis - one submission must have exactly one Target with a valid Type (Fv or BB Type - e.g. VHH, Mutein, etc.). In the Engineer Proteins Table, one can see the Sequence properties, Structure-derived properties, and any associated Experimental data in each of the relevant tabs on the top left side of the table for all entities in the Request matching the Target and Type.

Decision Matrix Column Selection

The Engineering Decision Matrix allows for a concise, mixed view of key columns which the user selects from any of the Sequence, Structure, or Experimental tabs. Columns are selected by checking the checkbox next to the column name in the table headers.

Predict & Analyze Structure

While sequence properties are calculated by default for all Fvs and BBs, structural property calculations must be initiated by the user. Once rows are selected, launch the structure prediction jobs by hitting the “Predict & Analyze Structure” button.

• Individual row - Select/deselect the checkbox on the far left side of the row of interest

• All rows - Select/deselect the top left checkbox in the header row

• Range of rows - Select one row (A), hold the Shift Key, and select another row (B) further down to select the range from A to B.

Status column represents the job status of the Predict & Analyze Structure event.

• Green badge - Complete

• Three loading bars - Job in progress

• Red Error Triangle - Error. Upon hover, a tooltip with the error message will appear. One can rerun the analysis by selecting the row again and hitting the “Predict & Analyze Structure” button again

• N/A - Job not yet triggered

The Analyze column either has “N/A” if no analysis exists, or a “PDB” button, which takes the user to the Engineering Pad for that row.

Filter Selection

In the Decision Matrix tab, the selected columns will appear under headers defining the source of the column (Sequence, Structure, or Experimental). Using the properties of interest, one can filter down the rows to those which optimize developability. Once rows are filtered and selected, one can proceed to either engineer the set of selected entities in the ePad (“Engineering Actions → “Engineer Selection”) or finalize the set of entities to save to the request (“Engineering Actions → "Finalize Engineering”).

The contents of the current table can be exported as a CSV or XLSX file by clicking the “Export table” button and selecting the file type. This button can be found in the expanded Settings/Filter section.

The Color Count column shows that row’s number of green, orange, and red flags for each property in the current table view.

The thresholds used to determine the property color assignment come from the Target Profile (see Decision Matrix section).

The Target Profile can be selected under the Settings/Filters section, which is expanded using the filter icon in the top right corner above the table.

The thresholds for the system-generated values in the Sequence and Structure Tables are under the Analysis Result Types tab in the Target Profile. The order of the rows, which can be rearranged by dragging the six dots on the left side of the panel, correspond to the order of the columns in the Sequence and Structure Tables.

Engineering Pad

The Engineering Pad provides a consolidated view of sequence and structural properties, the 3D Structure Viewer, and the Sequence Viewer. This enables the user to analyze entities from a structural perspective and engineer new variants directly on the sequence utilizing molecular properties mapped to the sequence and structure.

When the user hits “Engineer Selection” from the Decision Matrix, they are brought to the ePad with the selected entities loaded in the top left dropdown with the IDs listed next to their Alias and Generation #. The selected property columns from the Decision Matrix are shown by default in the Property Table - to toggle between selected properties and all properties, hit the funnel button . The property names are searchable in the search bar directly above the table and sortable by column.