InsilicoΣ
Drug Discovery, Cheminformatics & Bioinformatics
About Our Team Publications Mobile App Contact Us Login Register

Help & Tutorial

Learn how to use APPAS and interpret your results

How to Interpret Your Results

When your APPAS job completes, you will see a results page with several metrics and downloadable files. This section explains what each metric means and how to evaluate the quality of your polymer structure.

1. Overall Quality Score

The Overall Quality Score is a percentage (0–100%) that summarises how well the generated 3D structure passes all validation checks. It is computed from multiple weighted sub-metrics including bond lengths, steric clashes, valence, and connectivity.

Score Ranges
ScoreBadgeMeaning
90–100% Excellent Structure is suitable for molecular dynamics simulation and downstream analysis without modification.
70–89% Good Structure is usable but may have minor issues (e.g., a few steric clashes). Review the warnings.
<70% Needs Review Structure has significant issues. Check error messages and consider adjusting your input.
Tip

A score of 90%+ with all individual checks passed means your structure is ready for use in GROMACS, AMBER, or other MD simulation packages after force-field parameterisation.

2. Bond Length Metrics

Bond lengths are the most critical quality indicator. APPAS compares every bond in your polymer to element-specific ideal lengths from crystallographic databases.

Mean Bond Length

The average bond length across all bonds in the polymer. Typical values:

  • Polyethers (PEG, PPG): ~1.44–1.46 Å (dominated by C–O and C–C single bonds)
  • Polyesters (PLA, PGA, PCL): ~1.37–1.44 Å (shorter due to C=O double bonds)
  • Mixed copolymers: intermediate values depending on composition
Bond Length Std. Dev.

Measures how spread out bond lengths are around the mean.

  • Good < 0.10 Å — tight distribution
  • Caution 0.10–0.15 Å — some variation
  • Concern > 0.15 Å — indicates distorted bonds
% Bonds Within 5% of Ideal

Fraction of bonds whose lengths fall within 5% of the element-specific ideal. Higher is better.

  • >85% Excellent — most bonds near-perfect
  • 70–85% Good — acceptable for most uses
  • <70% Review structure carefully
% Bonds Within 10% of Ideal

The relaxed threshold. For a usable structure, this should be 100% (or very close).

  • 100% All bonds within acceptable range
  • 95–99% A few outlier bonds — check warnings
  • <95% Significant bond distortion detected
Key Reference: Ideal bond lengths are from Allen et al. (1987) crystallographic tables. C–C single = 1.54 Å, C–O single = 1.43 Å, C=O double = 1.23 Å, C–C double = 1.34 Å.
3. Steric Clash Count

Steric clashes occur when non-bonded atoms are too close together, creating unphysical overlaps. APPAS reports clashes at two distance thresholds:

Overlaps < 1.2 Å (Severe)

Atoms closer than 1.2 Å represent physically impossible overlaps. These must be zero for simulation-ready structures.

  • 0 Perfect — no severe clashes
  • 1–5 Minor — usually at chain junctions; often resolves after energy minimisation
  • >5 Review monomer SMILES or reduce chain length
Overlaps < 1.5 Å (Moderate)

Slightly less severe. Some moderate overlaps are expected in dense polyester regions due to short C=O bonds.

  • 0–5 Acceptable for most applications
  • 5–15 Common in polyesters; typically resolves with MD minimisation
  • >15 May indicate issues with monomer geometry
Note: Polyether-based polymers (PEG, PPG, Poloxamer) typically achieve zero clashes at both thresholds. Polyester-containing systems (PLA, PGA, PCL) may show a few moderate overlaps due to the shorter C=O bond geometry — this is normal and expected.
4. Molecular Weight (MW)

The molecular weight in Daltons (Da) is calculated from the exact atomic masses of all atoms in the polymer, including hydrogen atoms.

How to Verify

MW should approximately equal:
MW ≈ (monomer MW × repeat units) + end groups

For example, PEG with 20 repeat units:
44.05 × 20 ≈ 881 Da (plus end groups ≈ 1120 Da total)

What It's Used For
  • Verifying the polymer was built with the correct number of units
  • Input for pharmacokinetic models (drug delivery applications)
  • Estimating hydrodynamic properties
5. Radius of Gyration (Rg)

The radius of gyration measures the overall compactness of the polymer chain. It is the root-mean-square distance of all atoms from the centre of mass.

Interpretation
  • Smaller Rg → more compact, coiled chain
  • Larger Rg → more extended, stretched chain
Typical Ranges
Short chains (10–15 units)7–12 Å
Medium chains (15–25 units)10–15 Å
Long chains (25+ units)12–25 Å

Note: APPAS generates a single conformation in vacuum. In solution, Rg would differ due to solvent interactions.

6. End-to-End Distance (Ree)

The distance between the first and last atoms of the polymer chain. Together with Rg, it characterises the chain shape.

Ree / Rg Ratio
RatioShape
≈ 2.4 (√6)Random coil (ideal Gaussian chain)
≈ 3.5 (√12)Extended / rod-like conformation
< 2.0Compact / collapsed chain

Most APPAS-generated polymers fall in the 1.7–3.6 range, consistent with short chains in vacuum.

7. Output Files
FileFormatUse For
polymer.pdb PDB 3D visualisation (PyMOL, VMD, UCSF Chimera), MD simulation input. Contains CONECT records and multi-chain encoding.
polymer.mol MDL MOL Chemical structure editing (Avogadro, ChemDraw). Preserves bond orders and stereochemistry.
Properties CSV CSV Spreadsheet analysis. Contains all computed properties (MW, Rg, Ree, atom/bond counts).
Properties JSON JSON Programmatic access. Machine-readable property dictionary for scripting and data pipelines.
Analytics Excel XLSX Comprehensive analysis report with multiple sheets (summary, bonds, atoms, quality).

The PDB file uses multi-chain encoding where each monomer block is assigned a separate chain ID (A, B, C...), making it easy to analyse individual domains.

8. Quality Check Warnings & Errors

The quality panel on the results page may show Warnings and Errors. Here is what the common messages mean:

MessageSeverityWhat It MeansWhat to Do
Bond length deviation > 5% Warning Some bonds are slightly longer/shorter than ideal crystallographic values. Usually acceptable. If many bonds are flagged, try a longer monomer SMILES representation.
Steric clash detected Warning Non-bonded atoms are closer than 1.5 Å. A few clashes in polyesters are normal. For simulation, run a short energy minimisation first.
Bond length > 2.0 Å Error An unusually long bond was found — likely a disconnected fragment. Check your monomer SMILES for correct connectivity and exactly two * connection points.
Valence violation Error An atom has more bonds than its expected valence. Check your SMILES for duplicate bonds or incorrect atom types.
Disconnected fragments Error The polymer chain is not fully connected. Ensure each monomer SMILES has exactly two * (or [R]) connection points.

Step-by-Step Tutorial

Follow these steps to build your first polymer with APPAS. We'll use a PEG-PLA diblock copolymer as an example — one of the most widely used polymers in drug delivery.

1
Navigate to APPAS Dashboard

Go to the APPAS Dashboard and click "Create New Job". You will need to be logged in.

2
Define Your Monomers

You can either select monomers from the template library or enter custom SMILES. Each monomer must have exactly two connection points marked with * (asterisk) or [R].

Example Monomers
NameSMILESDescription
PEG*OCCO*Poly(ethylene glycol) — hydrophilic polyether
PLA*C(=O)OC(C)C(=O)O*Poly(lactic acid) — biodegradable polyester
PCL*OCCCCCC(=O)*Poly(caprolactone) — slow-degrading polyester
PGA*C(=O)OCC(=O)O*Poly(glycolic acid) — fast-degrading polyester
Important: Each SMILES must have exactly two * symbols. The first * is the "head" (connects to the previous monomer) and the second * is the "tail" (connects to the next monomer).
3
Build the Polymer Sequence

Add monomers to the sequence and specify the number of repeat units for each block. APPAS will connect them in order.

Example: PEG-PLA Diblock
OrderMonomerUnitsRole
1PEG10Hydrophilic block (chain A)
2PLA10Hydrophobic block (chain B)

The real-time molecular weight calculator at the bottom of the form shows the estimated MW as you adjust the repeat units.

4
Submit and Wait

Click "Build Polymer" to submit your job. APPAS will:

  1. Parse your monomer definitions and validate the SMILES
  2. Generate the 3D monomer structures using RDKit ETKDG embedding
  3. Connect monomers sequentially with dihedral sweep optimisation
  4. Resolve steric clashes using the push-apart algorithm
  5. Apply bond length corrections to ideal values
  6. Run the full validation suite
  7. Generate output files (PDB, MOL, properties)

Typical build time: 0.5–2 seconds for polymers with 10–30 repeat units. The page auto-refreshes every 5 seconds while processing.

5
Review Your Results

Once complete, the results page shows:

Quality Score

Overall structural quality as a percentage with colour-coded badge.

Properties Table

MW, atoms, bonds, Rg, Ree, and all computed metrics.

Download Files

PDB, MOL, CSV, JSON, and full archive (ZIP).

Refer to the "Interpreting Results" section above for detailed guidance on each metric.

6
Download and Use Your Structure

Choose the right file format for your purpose:

  • PDB → Load into PyMOL, VMD, or UCSF Chimera for 3D visualisation; use as input for GROMACS/AMBER after force-field parameterisation
  • MOL → Open in Avogadro or ChemDraw for chemical editing; preserves bond orders
  • CSV/JSON → Import into Python, R, or Excel for data analysis
  • Download All → Gets a ZIP archive with all files for archiving or sharing
7
Batch Processing (Advanced)

Need to build many polymers at once? Use the Batch Upload feature:

  1. Create multiple .txt input files (one per polymer) using the format below
  2. Package them into a single .zip file
  3. Go to Batch Upload and upload the ZIP
  4. APPAS processes all polymers and provides a combined report
Input File Format
# Comment lines start with #
PEG: *OCCO*
PLA: *C(=O)OC(C)C(=O)O*

[SEQUENCE]
10, PEG
10, PLA

FAQ & Tips

  1. Check SMILES syntax: Ensure each monomer SMILES is valid and has exactly two * connection points. You can verify SMILES at rdkit.org.
  2. Try longer SMILES: Very short SMILES (e.g., *OCC*) may fail RDKit 3D embedding. Use a fuller representation (e.g., *OCCO* for PEG).
  3. Reduce repeat units: Very large polymers (>100 units) take more time and memory. Start with 10–30 units.
  4. Check error message: The red error box on the results page provides the specific RDKit or APPAS error.

  • Use full ester linkage SMILES: For polyesters, include the full ester group in your SMILES. For example, use *C(=O)OC(C)C(=O)O* for PLA instead of *OC(C)C(=O)*.
  • Check monomer compatibility: When connecting different monomers (e.g., PEG + PLA), ensure the head/tail atoms form a chemically sensible bond.
  • The structure is still usable: APPAS applies bond length correction automatically. Even with some deviations, 100% of bonds typically fall within 10% of ideal values.

APPAS uses standard SMILES notation with connection points (dummy atoms) to define where monomers connect:

  • * — Asterisk marks a connection point (most common)
  • [R] — Explicit R-group notation (alternative)
  • [*] — Bracketed wildcard (equivalent to *)

Rules: Exactly two connection points per monomer. The first * is the "head" and the second is the "tail". APPAS connects the tail of monomer n to the head of monomer n+1.

The APPAS PDB file is GROMACS-compatible (single TER between chains, standard ATOM records, CONECT records with bond orders). However, you still need to:

  1. Generate force-field parameters using tools like ACPYPE, CGenFF, or GAFF (APPAS does not generate topology files)
  2. Run energy minimisation to relax any residual strain (typically 500–1000 steps of steepest descent)
  3. Add solvent and ions for solvated simulations

APPAS supports any linear polymer that can be defined from monomer SMILES with two connection points. This includes:

  • Homopolymers — single monomer repeated (e.g., PEG, PCL, PLA)
  • Diblock copolymers — two different blocks (e.g., PEG-PLA, PEG-PCL)
  • Triblock copolymers — three blocks, ABA or ABC (e.g., PEG-PPG-PEG, PEG-PHB-PHV)
  • Multi-block / statistical copolymers — any sequence of blocks
  • Novel monomers — any SMILES you can define, no pre-defined library required

A small number of steric clashes (<5 at the 1.2 Å threshold) is normal, especially for polyester-containing polymers. These typically occur at the junction between monomer blocks and resolve immediately with a short energy minimisation in your MD software.

If you see >10 severe clashes, consider: (1) using a longer monomer SMILES, (2) reducing the total chain length, or (3) checking that the head/tail connection chemistry is sensible.

AI Lab