Skip to content

Project View

The Project View provides a summary of all antibodies in the project.

[!NOTE] Multispecific Antibodies: Because multispecific antibodies consist of arbitrary multi-chain constructs (such as bispecifics or CrossMabs), they bypass sequence-level scoring and do not display residue-level stability, surface property, humanness, or liability analysis values in the Project View. For detailed information on designing and managing multispecific constructs, refer to the Assembler documentation.

To run standard analysis and scoring on the individual binding arms of a multispecific entry, you can use the Extract Fvs feature to isolate and save them as individual Fv entries in the project. For details, see the Extract Fvs documentation.

Rather than pulling all of these analyses together manually and building a speadsheet in order to rank order a discovery set and focus on lead selection and optimization, follow my core work principle and Be Inherently Lazy. Pull together your analyses into a single analysis tool and run the tool to do the hard work for you.

Full Project View

↔ Scroll horizontally to view full image

  • Sort & Filter: Click a header sort button to sort or its filter button to enter a filter value. Filtering is context sensitive and can filter by text or value ranges. Click the filter button again to clear a filter. Click the sort button again to sort in the opposite direction.
  • Name: The entry name associated with the analysis. Names may be double-clicked to be edited. If the entry name background is light red, it means that no PDB structure is available for that entry. If the cell contains a small eggshell square in the upper right corner, it means that the entry has engineering mutations.
    • If the entry is a Multispecific (has been through the Assembler), it will have a label 'Multispecific' and the analyses will not have been run. To see the individual analyses, use Extract Fvs to create individual Fv entries in the project.
    • If the entry light chain has a constant domain, and is missing or has an extra junction residue, the entry name will be flagged with a descriptive warning stating one of 'Missing R', 'Extra R', 'Missing G', 'Extra G'. Use Edit > Repair Junction to correct these issues and automatically re-run the analysis.
  • Lineage: The parent of the entry. If the entry is a variant, the parent is the antibody it was derived from. These may be set by the user, but are automatically set for Engineering builds.
  • Genetic Origin: The genetic origin and species of the antibody chains. This is determined as the closest germline sequence. These values are helpful for determining genetic diversity, comparison to known germline stabilities (but that can change with a single mutation), and known germline titer (but that can change with a single mutation). Non-human sequences are shown with a light orange background. Kappa VLs are showin in light Kobicha and Lambda VLs are shown in light Lavender.
  • Nearest Neighbor: The closest sequence to that entry within the set. This is a simple depiction of set diversity. Clading provides a more detailed view of set diversity.
    • Alignment: The residues for each antibody's variable light (VL) and variable heavy (VH) chains are aligned according to the selected numbering scheme (e.g., IMGT, Kabat, AHo, or Martin).
    • Merging: The aligned VL and VH positions are merged to represent the full variable fragment (Fv).
    • Distance Calculation: The sequence distance (p-distance) is calculated between every pair of antibodies in the project using the formula: Distance=100×(1−Percent Identity)
    • Comparable Length: The comparable length (denominator) is the sum of the positions present in the union of both sequences for both chains.
    • Identification: For each entry, the antibody that yields the lowest distance is identified as its Nearest Neighbor.
  • CDR Length: The measured sequence lengths, as well as the summed total, for the CDRs based on the defined region scheme.
  • Stability
    • AbLang, AbLang2, and IgBert: Summed LLM likelihood differences between the parent and most likely residues at each position for full-length sequence and framework positions only. AbLang is a broader LLM while AbLang2 and IgBert remove germline bias. Scoring is based on natural antibody distributions. See AbLang, AbLang2, and IgBert.
    • Disrupted CDR3 Salt Bridge: Identifies if the canonical HC-CDR3 salt bridge has been disrupted by mutation. This can impact fold stability.
  • Surface Properties
    These are calculated similar to the OPIG TAP method.
    • SPH: Hydrophobicity character of the surface.
    • SPP: Positive charge character of the surface.
    • SPN: Negative charge character of the surface.
    • SPCD: Fv charge interaction character of the surface.
  • Cysteines
    • Cysteine Missing: Identifies if a normal Ig fold cysteine is missing from the sequence.
    • Unusual Cysteine: Identifies cysteines in the sequence that are not in the normal Ig fold positions.
    • Unpaired Cysteine: Identifies unpaired cysteines in the sequence. After structure evaluation this count may be modified directly in the grid or in the Excel export to lower the score (lower is better). Performing a "Full Re-run Analysis" on the entry will reset the value.
  • Potential PTMs
    See LAP: Liability Antibody Profiler by sequence & structural mapping of natural and therapeutic antibodies. Detailed descriptions of these motifs are available in the Satlawa Liabilities Table.
    • Deamidation and Isomerization: If these are realized over time they could degrade antigen binding.
    • N-linked Glycosylation: Identifies potential glycosylation sites. While these could help solubilize the Fv, they may also impact affinity and efficacy if there are differences across productions.
    • Met and Trp Oxidation: Methionine and Tryptophan residues are susceptible to oxidation. If realized over time they could degrade antigen binding.
    • Hydrolysis: Creation from NP to DP during deamidation, which could then isomerize.
    • Fragmentation: Potential cleavage between defined residues.
  • Humanness
    A measure of the VL, VH, and Fv humanness of the sequence. OASign calculates the percentage of 9-mers in the sequence that are present in a curated database derived from the Observed Antibody Space (OAS). A higher score indicates a more human sequence.
  • Score
    The score of the entry, calculated based on the analyses performed. This value may be used for rank ordering the set. The function used in the score calculation is exposed in the Excel output. The values and weights used for the KBC score may be modified in the user settings section, as well as in the Excel export.
  • Notes
    User-editable notes for each entry. Double-click the cell in the grid to edit. Notes are saved to the database and persisted across analysis re-runs and engineering modifications. They are also included in the Excel and CSV exports.
  • Edit
    • Add Fvs and/or PDBs: Add entries to the project and/or add PDBs to existing entries.
    • Assembler: Open the Assembler interface for the selected Fvs.
    • Humanization: Open the Humanization interface for the selected Fv.
    • Engineering: Open the Engineering interface for the selected Fv.
    • Engineer pI: Open the pH interface for the selected Fv.
    • Set Parent: Set or clear the parent of the selected entries.
    • Extract Fvs: Extract variable domains (Fv pairs) from the selected antibodies (such as full-length IgG or multispecific formats). Users can designate numbered pair suffix assignments (up to 10 pairs) to separate multiple VH/VL chains, which are then saved as new individual Fv entries in the project for humanization, engineering, or assembly.
    • PDB Files: Open the PDB Files interface. This allows for the downloading or deletion of PDB files. If a PDB file is deleted the entry must be re-run to remove its surface property values. While AbLead could be used as a PDB repository, try JStruct instead. Or, for a commercial product, try Molecular Operating Environment's PSILO.
    • Build Models: Generate 3D structural Fv or VHH models for the selected entries using ABodyBuilder2 or NanoBodyBuilder2. This process calculates the variable domain structure and automatically updates the entry with its corresponding surface properties.
      • Note: This builds only the variable domain models, which might not give accurate surface properties since residues which would be buried by the CL/CH domains in a Fab structure will be exposed to solvent in the Fv model.
      • NOTE!: Verify the structure models external to AbLead before relying on their validity! Use the PDB Files interface to download and inspect the models.
    • Delete Selected: Delete the selected entries. These are deleted from the project and the database immediately.
  • Export
    • Excel: Export the selected entries to an Excel file with conditional formatting and a weighted function to generate the scores.
      The attributes used to generate the score, each with their own weight, are: L1, L2, L3, H1, H2, H3, CDR Sum, Salt Bridge, Surface Properties (SPH, SPP, SPN, SPCD), Liabilities (Cysteines, Deamidation, Isomerization, N-glycosylation, Oxidation, Hydrolysis, Fragmentation), AbLang/AbLang2/IgBert (FR Score), Humanness Score (OASign VL, VH, Fv), and pI (Bjellqvist).
    • CSV: Export the selected entries to a CSV file.
    • Analysis Report: Export the selected entries to a Word document analysis report for further editing.
    • Sequences: Visualize or export the selected entries' sequences in multiple formats.
    • WIPO ST.26 (Patent)...: Create a WIPO ST.26 (Patent Sequence Listing) formatted sequence data file for patent applications.
  • Share: Open the sharing dialog to grant or manage access to the project for other users within the same organization (same email domain) (requires appropriate permissions).

Share

References