Skip to main content

Diagnostics

The Diagnostics page provides low-level chemical structure analysis tools for debugging parsing errors, investigating structural discrepancies, and pre-screening files before batch processing.

Access Diagnostics from the Data Preparation dropdown in the header navigation.

Available Tools​

SMILES Diagnostics​

Parse and analyze a SMILES string to understand its structure at the atom level:

  • Atom-by-atom breakdown: Each atom with its element, charge, isotope, and bonding
  • Ring closures: Identification of ring opening/closing tokens
  • Branch points: Parenthetical branch structure
  • Parse errors: Detailed error messages for invalid SMILES with the position of failure
Debugging Invalid SMILES

When a molecule fails validation with "Cannot parse input," paste it into SMILES Diagnostics to see exactly where and why parsing fails.

InChI Layer Diff​

Compare InChI layers between two molecules to identify exact structural differences:

  • Formula layer — Molecular formula comparison
  • Connectivity layer — Atom connectivity differences
  • Hydrogen layer — Hydrogen attachment differences
  • Charge layer — Formal charge differences
  • Stereo layers — Stereochemistry differences (tetrahedral, double bond)

This is useful for understanding why two molecules that look similar have different InChIKeys or why standardization changed a structure.

Round-Trip Validation​

Validate SMILES → MOL → SMILES round-trip fidelity:

  1. Parse the input SMILES into an RDKit molecule object
  2. Convert to MOL block (3D coordinates if possible)
  3. Convert back to SMILES
  4. Compare the original and round-tripped SMILES

Differences indicate potential representation loss — the molecule may not survive serialization/deserialization cycles in some workflows.

File Pre-Validation​

Check a file for parseable SMILES before uploading to batch processing:

  • File parsing: Validates that the file format is correct (CSV, SDF)
  • Column detection: Identifies SMILES and name columns
  • Parse rate: Reports what percentage of SMILES can be parsed by RDKit
  • Error summary: Lists unparseable entries with row numbers and error messages
Pre-Screening Saves Time

Running file pre-validation before a large batch upload helps you fix format issues and unparseable entries upfront, rather than discovering them during processing.

Coordinate Dimension Analysis​

Detect the coordinate dimensionality of molecules in MOL blocks or SDF files:

DimensionMeaning
2DX/Y coordinates only (Z = 0) — suitable for 2D depiction
3DX/Y/Z coordinates — includes conformational information
No coordinatesNo atom coordinate data present
MixedFile contains molecules with different dimensionalities

This helps verify that SDF files have the expected coordinate data before using them for 3D analysis or docking workflows.

Use Cases​

Debugging Validation Failures​

  1. A molecule fails validation → copy the SMILES
  2. Open Diagnostics → SMILES Diagnostics
  3. Paste the SMILES → see the exact parsing error or structural issue
  4. Fix the SMILES and re-validate

Comparing Standardized vs. Original​

  1. Standardize a molecule → note the original and standardized SMILES
  2. Open Diagnostics → InChI Layer Diff
  3. Enter both SMILES → see exactly which structural layers changed
  4. Understand whether stereochemistry, connectivity, or charge was affected

Pre-Screening Batch Files​

  1. Open Diagnostics → File Pre-Validation
  2. Upload your CSV or SDF
  3. Review the parse rate and error summary
  4. Fix problematic entries
  5. Upload the cleaned file to Batch Processing

Next Steps​