R&D SDS · CRO & Pharma

SDS for Novel Compounds: How CROs Generate Safety Data Sheets for Proprietary APIs

Your novel intermediate or proprietary API candidate doesn't exist in PubChem, ECHA CHEM, or any standard SDS database. Your safety officer still needs documentation before the compound enters the lab. This guide explains exactly how contract research organizations and pharma labs solve this problem — and what makes a proxy SDS scientifically defensible.

Why Standard SDS Databases Fail for Novel Compounds

Every major SDS database — PubChem, ECHA's Chemical Database, ChemSpider, ChemWatch, ICSC — is built around registered, publicly-known substances. A chemical entry in these databases requires a CAS Registry Number, which is only assigned after a substance is formally registered with the CAS Division of the American Chemical Society.

Novel compounds synthesized in a CRO or pharma discovery lab are, by definition, not yet in any public registry. They may be:

Synthetic intermediates

Compounds synthesized on the route to a target molecule — not yet characterized, not yet named

API candidates (NCEs)

New Chemical Entities in early discovery — lead compounds before IND filing

Proprietary formulations

Mixtures or derivatives under NDA whose composition is a trade secret

Analogue libraries

Sets of closely related compounds synthesized for SAR screening — dozens may be made per week

Pre-candidate compounds

Molecules selected for further optimization before formal registration

Undisclosed SMILES

Compounds identified only by internal code (e.g., Cpd-23B, NVX-301) without public identifiers

When you enter any of these into a standard SDS generator, you receive: no results. The lookup fails because there is genuinely nothing to find. The compound has no CAS, no PubChem CID, no ECHA CHEM entry. You can't download what doesn't exist.

The compliance gap this creates

Under OSHA's Hazard Communication Standard (29 CFR 1910.1200) and EU REACH Annex II, employees handling hazardous chemicals must have access to an SDS. The regulation doesn't exempt novel compounds — it just creates a documentation vacuum that CROs and pharma labs must fill themselves.

Traditional solutions — commissioning a regulatory consultant, waiting for CAS assignment, or skipping documentation entirely — are either prohibitively expensive, too slow for active discovery programs, or non-compliant. There is a better approach.

How R&D SDS Generation Works

The scientifically validated approach to generating an SDS for a novel compound is called analogue-based extrapolation, or read-across. It is not guesswork — it is the same methodology ECHA formally documents in its Read-Across Assessment Framework (RAAF), used throughout REACH substance evaluation.

The core premise: structurally similar compounds tend to exhibit similar physicochemical, toxicological, and ecotoxicological properties. If we can find a well-characterized compound that is structurally very close to your novel molecule, we can conservatively bridge its safety data to populate a proxy SDS.

Step 1 — Structural Similarity Search

The process begins with your compound's structural representation. The preferred input is a SMILES string(Simplified Molecular Input Line Entry System), which fully encodes your compound's molecular graph without disclosing a registered name. InChI and InChIKey are also accepted and converted to SMILES server-side via RDKit.

The system searches ECHA's chemical database for the closest structural match using the Tanimoto coefficient — a mathematical measure of structural overlap ranging from 0 (completely dissimilar) to 1 (identical). Only analogs scoring Tanimoto ≥ 0.85 are used as proxy sources. This threshold aligns with ECHA RAAF requirements for high-confidence structural analogy.

Tanimoto Coefficient Scale

≥ 0.90Very high similarity — core scaffold and key substituents largely identical
0.85–0.89High similarity — qualified threshold for ECHA RAAF read-across
0.70–0.84Moderate similarity — below threshold, no proxy SDS generated
< 0.70Low similarity — structurally distinct, read-across not defensible

Step 2 — Read-Across Data Bridging

Once a qualifying analog is identified, hazard data is carried across to populate all 16 GHS sections of the proxy SDS. This includes:

  • GHS hazard classification — Based on analog's confirmed classification — conservative where structural differences introduce uncertainty
  • Physical & chemical properties — Molecular weight, estimated logP, predicted boiling/melting points from SMILES-derived calculations
  • Acute toxicity (Sections 11) — LD₅₀/LC₅₀ values from analog, flagged as estimated where extrapolated
  • Ecotoxicological data (Section 12) — Aquatic toxicity and persistence data from analog
  • Transport classification (Section 14) — UN number and packing group assigned based on hazard classification

Step 3 — Confidence Disclosure

Every section of the proxy SDS includes a confidence label: confirmed data (from the analog), estimated data (extrapolated with conservative assumptions), or not available. The proxy source compound, its CAS number, Tanimoto score, and source database are disclosed in Section 1 and in a dedicated proxy disclosure appendix.

This transparency is not optional — it is what makes the document defensible. Your safety officer can review exactly which data is confirmed and which is estimated, and apply appropriate additional precautions where uncertainty exists.

R&D SDS Tool

Try a Free Demo SDS for Your Novel Compound

Submit a SMILES string or compound name and see a complete 16-section proxy SDS — watermarked for evaluation, no subscription required.

Try Free Demo SDS — from $49

1 demo per IP every 24 hours · Subscription plans from $299/mo for active pipelines

REACH and GHS Compliance for Research-Stage Compounds

Compliance obligations for novel compounds in R&D are nuanced — and frequently misunderstood. Here is the accurate regulatory picture:

OSHA HazCom 2012 (US — 29 CFR 1910.1200)

Requires that employees working with hazardous chemicals have access to SDS. For novel compounds with incomplete toxicological characterization, OSHA requires that SDS contain “all available and reasonably ascertainable information” — meaning a conservatively-estimated proxy SDS satisfies the requirement better than no documentation at all.

GHS Section 1 of the proxy SDS must clearly identify it as a proxy/estimated document for R&D use only, with the generating organization and the disclaimer that it is not for commercial distribution.

EU REACH (Regulation EC 1907/2006)

REACH Article 2(1)(d) provides an R&D exemption for substances used in scientific research and development. Pre-candidate compounds handled exclusively within controlled laboratory conditions are generally exempt from REACH registration during the R&D phase.

However, REACH Annex II still applies to any substance supplied to another party— including CRO supply. If your pharma organization sends a novel compound to a contract lab, the accompanying SDS must conform to Annex II format. A proxy SDS with proper labeling satisfies this requirement for R&D quantities.

EU CLP (Regulation EC 1272/2008)

CLP requires hazard classification and labeling for substances placed on the EU market. For R&D quantities not placed on the market (internal handling only), full CLP classification is not mandatory. A proxy SDS aligned with GHS Rev 7 classification criteria — as generated by ChemEngine — is appropriate for internal lab handling and CRO supply documentation.

Important scope limitation

Proxy SDS documents are appropriate for internal R&D use, CRO supply documentation, and laboratory safety compliance. They are not suitable for REACH registration submissions, commercial product distribution, transport of commercial quantities, or any regulatory filing that requires confirmed experimental data.

Use Cases: Who Needs R&D SDS?

The need for novel compound SDS documentation spans multiple roles and organizations in the drug discovery and specialty chemicals pipeline:

🏭

Contract Research Organizations (CROs)

CROs receive novel compounds from pharma clients for synthesis, testing, or scale-up — often under NDA and without any public registration. Safety documentation must accompany these compounds on arrival to satisfy OSHA HazCom requirements and internal EHS protocols.

Volume problem: Active CROs may receive dozens of novel compounds per week, each requiring documentation before lab work can begin. Manual SDS authoring at $2,000–$8,000 per document is economically impossible at scale.

💊

Pharma Medicinal Chemistry Teams

In drug discovery, medicinal chemists synthesize and screen hundreds of analogue compounds per program in search of lead candidates. A typical SAR campaign generates 50–200 novel compounds before a lead is identified — almost none of which will proceed far enough to justify a traditional SDS.

Typical workflow: Batch generate proxy SDS for entire analogue libraries via SMILES-based bulk upload. Update documentation as candidates are deprioritized or promoted to the next stage.

🔬

Academic Spin-Outs & Biotech Startups

Early-stage biotechs and academic spin-outs often lack the regulatory infrastructure of large pharma organizations. They need GHS-compliant documentation for investor due diligence, CRO supply, and internal safety protocols — but typically have neither the budget for a dedicated EHS function nor the throughput to justify enterprise SDS authoring software.

Key need: Audit-ready documentation that satisfies both OSHA HazCom and EU REACH Annex II requirements for early-stage lab use and CRO supply.

⚗️

Specialty & Fine Chemical Producers

Companies synthesizing novel specialty chemicals, process intermediates, or custom reagents for industrial customers face the same database gap. A new synthetic intermediate used in multi-step manufacturing may never be registered publicly — but must still have safety documentation for employees handling it.

Ongoing use: Proxy SDS can be used as working documentation while the substance is being characterized for formal registration, then replaced with a fully confirmed SDS once registration is complete.

Try It Now

Generate a Proxy SDS for Your Novel Compound

ChemEngine R&D SDS: 16 GHS sections, ECHA RAAF methodology, Tanimoto ≥ 0.85, editable .docx with full confidence disclosure. Subscription plans from $299/mo.

Try Free Demo SDSView Subscription Plans

Starter $299/mo · Pro $999/mo · Enterprise $2,499/mo · No setup fee · Cancel anytime

Frequently Asked Questions

Can I generate an SDS for a compound with no CAS number?
Yes — and this is exactly the use case R&D SDS tools are designed for. Standard databases (PubChem, ECHA CHEM, ChemSpider) index only registered, publicly-known substances. Novel compounds synthesized in your lab won't have a CAS number or any database entry. R&D SDS generation works by accepting a structural descriptor — typically a SMILES string or InChI — and finding the closest structurally similar compound that does have documented safety data. Hazard properties are then bridged across using the ECHA Read-Across Assessment Framework (RAAF). The resulting proxy SDS covers all 16 GHS sections and clearly labels which data was estimated vs. confirmed.
What is analogue-based extrapolation for SDS generation?
Analogue-based extrapolation — also called read-across — is a methodology in which hazard data for an untested (source) substance is predicted using data from one or more structurally similar, well-characterized (target) substances. The scientific basis is that structurally similar compounds tend to have similar toxicological and physicochemical properties. ECHA's Read-Across Assessment Framework (RAAF) provides documented criteria for when read-across is scientifically defensible: the analogue must be structurally similar (typically quantified using a Tanimoto coefficient), the rationale for similarity must be documented, and uncertainties must be disclosed. For R&D SDS generation, we require Tanimoto ≥ 0.85 to generate a document — if no qualifying analog is found, no SDS is produced.
What is the OSHA requirement for novel compounds in the lab?
Under OSHA's Hazard Communication Standard (29 CFR 1910.1200), chemical manufacturers, importers, and distributors must produce SDS for hazardous chemicals. However, the Standard also recognizes that for research-stage, new chemical entities (NCEs) that haven't been fully characterized, documentation must still be provided to the extent information is reasonably available. The practical interpretation: your safety officer must have some documented hazard assessment before the compound is handled — even if that assessment is a conservative proxy estimate derived from structural analogues. A properly labeled proxy SDS generated via documented methodology satisfies this requirement for internal R&D use.
Is a proxy SDS sufficient for REACH compliance in the EU?
For substances handled exclusively in research and development (R&D), the EU REACH Regulation (EC 1907/2006) provides a specific exemption (Article 2(1)(d)) for substances used in scientific R&D. This means many pre-candidate compounds are not subject to full REACH registration obligations during the R&D phase. However, REACH Annex II still requires that an SDS accompany any hazardous substance or mixture placed on the market — even R&D quantities supplied to a contract lab. A proxy SDS clearly labeled as R&D use only, generated with documented methodology, is appropriate documentation for internal lab handling and for CRO supply chain use. It is not sufficient for commercial supply, REACH registration submissions, or transport of commercial quantities.
How long does it take to generate an SDS for a novel compound?
With ChemEngine's R&D SDS service, generation takes under 10 minutes from compound submission to .docx delivery. You submit your compound (SMILES string, InChI, InChIKey, compound name, or partial CAS), our system performs a Tanimoto similarity search against ECHA's chemical database, selects the best qualifying analog (≥0.85 threshold), applies read-across hazard data across all 16 GHS sections, and delivers an editable proxy SDS with full confidence disclosure. By comparison, commissioning a regulatory consultant to write a custom SDS for a novel compound typically takes 3–6 weeks and costs $2,000–$8,000 — a timeline incompatible with active discovery programs.
What input formats are accepted for novel compound SDS generation?
ChemEngine's R&D SDS generator accepts five input types: (1) SMILES string — strongly preferred, enables the most accurate Tanimoto similarity scoring; (2) InChI or InChIKey — converted to SMILES server-side via RDKit; (3) Partial or non-standard CAS — used as fallback text search in PubChem; (4) Compound name or internal designation (e.g., 'Cpd-23B', 'NVX-301') — triggers a fuzzy name search; (5) Molecular formula — lowest precision, may return multiple candidate analogs for selection. For proprietary compounds under NDA, SMILES is the preferred format because it fully encodes structural information without revealing a registered name.