Catalexis Reaction Optimization Platform FAQs

What comes in the Catalexis kit?

Catalexis is available in two formats:

Option A: 50 mg Ligand Kit (via our ordering platform AMS):

1. 23 phosphine ligands × 50 mg each
2. 500 mg Pd₂(dba)₃ palladium source
3. 5 Catalexis portal tokens (~5 complete screening runs, 115 reactions total)

Option B: 1 µmol Screening Plates:

1. Five 24-well reaction screening plates, each well with 1 µmol of ligand in 7x21mm 50 µL vials
2. Pd₂(dba)₃ supplied separately in a glass vial
3. 5 Catalexis portal tokens (optimized for automated HTE workflows)

Both formats include 5 portal tokens (1 token = 1 ligand ranking report). Additional tokens and custom kit configurations are available on request. Submit a quote request via the Catalexis Request Form.

Can I purchase individual ligands or related products?

Yes. All 23 phosphine ligands in the kit are available individually on our website. Ligand ranking reports include product numbers and direct ordering links for seamless reordering. Key ligands:

• DavePhos, Cat. No. 638021
• XPhos, Cat. No. 638064
• JohnPhos, Cat. No. 638439
• APhos, Cat. No. 677264
• tBuBrettPhos, Cat. No. 730998
• HandaPhos, Cat. No. 799580

The full catalog list is available on the Catalexis service page.

Note: while the 23 kit ligands are all available through our catalog, not all ligands recommended by the AI ranking engine (from the broader 400+ ligand database) are available for purchase through our catalog. For ligands outside our catalog, your account manager can advise on alternative sourcing options or contact Catalexis team at Catalexis@sigmaaldrich.com.

How does the workflow operate end-to-end?

Catalexis operates in two phases. Phase 1 covers ligand screening for Pd cross-coupling reactions. Phase 2 is the BayBE-powered reaction optimization campaign, which can be used for any reaction class, independently of Phase 1.

Phase 1: Ligand Screening (for Pd cross-coupling reactions)

1. Step 1: Order your kit, choose the 50 mg Ligand Kit or the 1 µmol Screening Plates.
2. Step 2: Run your 23-ligand screen using your chosen substrate and conditions. Keep all variables constant across all 23 reactions; vary only the phosphine ligand. Compatible with manual benchtop and automated HTE setups.
3. Step 3: Upload results to the Catalexis portal. Enter 23 integer yield values (0-100). No structural data, SMILES strings, or chemistry metadata required. Redeem 1 token.
4. Step 4: Receive AI-driven ligand rankings. Catalexis compares your data against its database of 400+ phosphine ligands and returns the top 25 candidates with product numbers and direct links.

Phase 2: Reaction Optimization (BayBE Campaign)

1. Step 5: Define your reaction space. Specify variables such as ligand, metal source, base, solvent, temperature, and equivalents. Custom variables can be added.The design space supports up to 1,000,000 reaction combinations.
2. Step 6: Generate BayBE recommendations. The Bayesian ML engine proposes the most information-rich experiments to run next, balancing exploration with exploitation. You control the batch size (number of experiments per batch) to perfectly fit your existing workflow.
3. Step 7: Run the recommended experiments and upload your results to receive the next batch of recommendations.
4. Step 8: Repeat Steps 6-7 until your reaction achieves the target yield, conversion, selectivity, or custom metric.

Phase 2 can be used independently of Phase 1 for any chemical reaction, not just Pd-catalyzed cross-coupling. Portal-only access (without purchasing a kit) is available for Phase 2 campaigns.

What data do I upload? Do I need to share structures or mechanisms?

No structural or mechanistic data is required. For Phase 1 (ligand screening), you only enter 23 integer values (0-100) representing observed reaction yields. No substrate structures, SMILES strings, or chemistry metadata are needed.

For Phase 2 (BayBE optimization), you define the reaction variables (e.g., ligand identity, base, solvent, temperature), but no structural information about your substrate or target molecule is required.

This design fully protects your intellectual property. Compound structures and reaction targets remain entirely confidential.

What optimization approach does Catalexis use?

Catalexis uses BayBE (Bayesian Back-End), an internally developed Bayesian machine learning optimization framework. Unlike fixed Design of Experiments (DOE), BayBE:

1. Builds a probabilistic surrogate model of your reaction from completed experiments.
2. Identifies the most informative experiments to run next, balancing exploration of unknown regions with exploitation of high-performing conditions.
3. Updates after each iteration, becoming progressively more accurate.

Fewer experiments are needed to converge on optimal conditions, even for complex, high-dimensional spaces with dozens of variables. The default target is yielding maximization; selectivity or byproduct minimization can also be configured.

Which reactions does Catalexis best support today?

The 23-ligand phosphine screening kit (Phase 1) is designed for Pd-catalyzed cross-coupling reactions using monophosphine ligands. Best-validated reaction classes:

1. Buchwald-Hartwig aminations (C-N bond formation)
2. Suzuki-Miyaura couplings (C-C bond formation via boronic acids)

Other Pd-catalyzed cross-couplings sensitive to phosphine ligand choice are also well-suited.

The BayBE optimization engine (Phase 2) is reaction-agnostic and can be applied to any chemical transformation where you can define variables and measure an outcome.

Is Catalexis suitable for HTE and manual workflows?

Yes, Catalexis supports both:

1. Manual benchtop setups: The 50 mg Ligand Kit is ideal for labs running individual reactions at the bench with standard HPLC, GC, or NMR analytics.
2. Automated HTE workflows: The 1 µmol Screening Plates format is configured for high-throughput platforms, minimizing weighing and handling errors in miniaturized experiments.

The portal accepts yield data regardless of how experiments were run, making integration into either workflow seamless.

Can I use Catalexis in route scouting or late-stage optimization?

Yes, Catalexis applies across the development continuum:

1. Route scouting / early development: Rapidly identify which ligand class enables a key coupling step before committing to extensive reagent development.
2. Process development: Use Phase 2 (BayBE) to optimize conditions (base, solvent, temperature, equivalents) at process scale.
3. Late-stage optimization: One customer identified the optimal ligand for an industrial-scale di-amination process after just three rounds of Catalexis screening.

What kind of improvements can I expect?

Compared to traditional catalyst screening, Catalexis delivers improvements across multiple dimensions:

1. Speed: Optimization in days rather than weeks using AI-guided experiment selection.
2. Ligand coverage: Effective screening of 400+ ligands from just 23 physical reactions, eliminating the constraint of "only screening what's on the shelf."
3. Material efficiency: Significantly less catalyst, substrate, and solvent consumed due to fewer total experiments.
4. Experiment count: Bayesian models select only high-value runs, redundant or uninformative experiments are avoided.
5. Discovery: Users consistently identify high-performing ligands they would not have independently considered.

Note: Specific quantitative benchmarks (average yield uplift, experiments saved) can be requested from the Catalexis team via Catalexis@sigmaaldrich.com.

How many iterations are typically needed?

The number of iterations depends on the complexity of the chemistry and the optimization target. Based on user experience:

1. Ligand identification (Phase 1): A single 23-ligand screen plus one or two BayBE rounds often identifies high-performing candidates.
2. Full process condition optimization (Phase 2): Typically 3-5 iterative rounds for moderately complex parameter spaces. The optimization engine generally gives good results from sampling 5-10% of the overall reaction space.

Note: For guidance on typical iteration counts for your reaction type, contact Catalexis@sigmaaldrich.com.

How do I design the first 23-ligand screen?

The key principle is controlled variation: choose your substrate, coupling partner, base, solvent, temperature, and Pd loading based on standard conditions for your reaction class, then vary only the phosphine ligand across all 23 reactions. Consistent conditions ensure yield differences are attributable to the ligand, giving the model clean, interpretable signal.

Use the Pd₂(dba)₃ supplied in the kit as the Pd source and apply the appropriate ligand:Pd ratio for your reaction. Each ligand is supplied at a quantity supporting approximately 5 screening reactions.

What if my analytical method gives conversion, not isolated yield?

The Catalexis portal accepts any numeric metric scaled 0 to 100. If your analytical method reports conversion (e.g., by HPLC area%, GC%, or NMR) rather than isolated yield, you can enter these values directly. Consistency across all 23 reactions is what matters, the model will optimize against whichever metric you report.

In Phase 2 (BayBE optimization), you can define custom optimization targets such as conversion, selectivity ratio, or byproduct suppression.

Note: For method-specific guidance on normalizing analytical outputs, contact Catalexis@sigmaaldrich.com.

Can I seed the model with prior internal data?

For Phase 2 (BayBE optimization campaigns), previously generated internal data can be incorporated as a starting point, giving BayBE a head start and reducing the number of additional experiments required.

Note: The specific mechanism for uploading prior data (file format, column requirements) should be confirmed with the Catalexis team at Catalexis@sigmaaldrich.com.

What if several ligands tie at similar yields?

When multiple ligands produce comparable yields in Part 1, include all high-performers in your Phase 2 BayBE optimization campaign as discrete variables. The BayBE engine will differentiate between them by exploring other parameters (base, solvent, temperature), often revealing one ligand outperforms under specific condition combinations.

Testing at least four of the top-ranked ligands as starting points for the optimization phase is generally recommended to ensure robustness.

Does the portal support batch records or audit trails?

Batch record and audit trail functionality, including whether uploads, edits, and token redemptions are logged, should be confirmed directly with the Catalexis team. This is important for GxP-adjacent workflows. 

What sample size, analytical method and QC checks are recommended for reliable 0-100 inputs?

While Catalexis accepts any yield/conversion values from 0-100, result quality improves with reliable analytics. General best practices:

1. Use a consistent, quantitative analytical method across all 23 reactions (e.g., HPLC with internal standard, GC-FID with calibration, or quantitative NMR).
2. Run at least one replicate for reactions yielding unusually high or low values to confirm they are not outliers.
3. Include a positive control (a known working ligand from prior experience) and a negative control (no ligand) where feasible.
4. Round to the nearest integer for portal entry.

How should I handle outliers or plate failures before uploading?

If a well fails (vial leak, evaporation, analytical error), do not enter 0, this could be misinterpreted as the ligand producing no product. Recommended options:

1. Repeat the failed reaction before uploading, if time allows.
2. Flag the failed well as a missing value before submission, check the portal interface for how to handle missing data points.

What are best practices for a second-round screen when several ligands tie?

When moving into a second round with tied ligands, consider:

1. Including all tied ligands in a BayBE campaign alongside at least one other variable (base or solvent) to provide differentiation.
2. Running confirmatory duplicates on tied ligands under original conditions to rule out experimental variance.
3. Setting selectivity or byproduct suppression as a secondary optimization criterion if yield alone does not differentiate candidates.

How does the Catalexis algorithm interpret my 23-ligand screening data?

Catalexis uses your 23 experimental yield values as a training signal to learn which structural or electronic features of phosphine ligands correlate with success in your specific reaction. It then maps this learned pattern across its database of 400+ phosphine ligands to estimate likely performance, surfacing candidates not included in the original screening plate that are predicted to perform well.

The system does not require any chemical knowledge about your substrate, only the yield values and the identities of the 23 screened ligands (fixed and known from the kit composition).

Does Catalexis build a mechanistic model of my reaction?

No. Catalexis is a data-driven, empirical model, it does not model reaction mechanisms, transition states, or electronic structure. It learns statistical patterns from your yield data and ligand property descriptors.

This approach is intentionally substrate-agnostic and privacy-preserving, mechanistic modeling would require detailed structural information about your compounds, which Catalexis does not collect.

How does Catalexis handle noisy or variable data?

BayBE is inherently designed to handle experimental noise and uncertainty. Rather than treating each data point as a perfect measurement, the Bayesian model maintains a probability distribution over possible outcomes, accounting for measurement variability when making recommendations.

In practice, the model is more robust than traditional DOE to occasional analytical errors or run-to-run variability. However, large systematic errors (e.g., consistent analytical bias, plate failures) should be addressed before uploading to avoid misleading the model.

Will the model overfit my small dataset?

Bayesian optimization is specifically designed to perform well on small datasets, this is one of its core advantages over classical ML approaches. Because the model encodes uncertainty explicitly, it naturally avoids overconfident predictions from limited data. Its acquisition function balances exploration with exploitation to avoid local optima.

The 23-ligand dataset is sufficient for meaningful initial predictions; accuracy improves with each subsequent iteration as more data is incorporated.

Can Catalexis predict ligands not included in the 23-ligand kit?

Yes, this is a core feature. The 23-ligand kit provides a diverse, information-rich training set. Catalexis uses this to extrapolate predictions across its database of 400+ phosphine ligands, ranking all of them by predicted suitability. The top 20 predictions, including ligands not in the screening plate, are returned with product numbers and direct links.

Does Catalexis explore reaction condition variables (temperature, base, solvent)?

Yes, through the BayBE optimization campaign (Phase 2). You define the full parameter space to explore, which can include any combination of:

1. Ligand identity, metal source, and catalyst loading
2. Base identity and equivalents
3. Solvent (single or mixed) and concentration
4. Temperature and reaction time
5. Substrate equivalents and any custom variables

The BayBE engine recommends which combinations to test first to maximize information gain. The design space supports up to approximately 1,000,000 reaction combinations.

Does Catalexis integrate computational chemistry descriptors (steric/electronic parameters)?

The internal ligand-ranking model uses encoded ligand property information (which may include steric and electronic descriptors) to map yield patterns across the 400+ ligand database. The platform does not currently accept user-supplied descriptors (e.g., from DFT calculations or Tolman cone angle databases) as direct inputs.

Can I export prediction data for in-house computational analysis (QSAR, DFT screening)?

Catalexis supports CSV export of results and recommendations, which can be used as input for downstream analysis. Whether full ranking scores, model confidence intervals, or ligand descriptors are included in the export depends on the current portal version.

Note: Confirm the exact content of exported files and whether JSON or other structured formats.

Will ligand-rank predictions remain valid when scaling from µmol to gram scale?

Ligand rankings from Part 1 are derived from miniaturized screening data and are most directly applicable at that scale. Catalyst performance can shift at larger scales due to mass transfer, heat effects, mixing, and concentration differences.

Catalexis recommends testing at least four of the top-ranked ligands at your target scale before committing to a final process ligand. Phase 2 (BayBE) can be used to confirm and fine-tune conditions at the scale of interest.

Does the model adjust predictions after each new iteration?

Yes. After each round of experiments, you upload your new results and BayBE updates its surrogate model with the additional data. Each iteration produces progressively more refined recommendations as the model becomes more confident about which regions of the chemical space perform well. This is the core strength of Bayesian optimization over fixed DOE designs.

Can Catalexis model multi-substrate datasets (e.g., substrate generality studies)?

Multi-substrate modeling (using yield data from several substrates to build a generality-aware model) is an advanced use case not explicitly described in current platform documentation. Please confirm with the Catalexis team whether this is supported or planned for a future release.

Can the platform perform virtual screening without running the initial physical 23-ligand screen?

Phase 1 ligand ranking engine requires physical experimental data from the 23-ligand screen, it cannot rank ligands without at least some experimental signal from your specific reaction system. Virtual-only prediction without any wet chemistry data is not currently supported for Phase 1.

For Phase 2 (BayBE optimization), if you already have historical data for your reaction, you may be able to start without a new physical screen. Portal-only access (without a kit) is available for Phase 2.

Does Catalexis consider solvent-ligand or base-ligand compatibility effects computationally?

The BayBE engine can explore interactions between solvent, base, and ligand simultaneously through the experimental design, it will discover compatibility effects from your data. Catalexis does not pre-encode known chemical incompatibilities as constraints in the model by default.

Note: Confirm with the Catalexis team at Catalexis@sigmaaldrich.com whether known incompatibility can be set as exclusions during reaction parameter space setup.

How do I de-risk translation of ranked ligands from µmol to gram scale?

Recommended approach for de-risking scale-up:

1. Test at least four top-ranked ligands under intended scale conditions before selecting a final candidate.
2. Use Phase 2 (BayBE) to optimize conditions at the target scale, adjusting catalyst loading, temperature, solvent volume, and mixing.
3. Run robustness experiments on your shortlisted ligand by sweeping key parameters (±10°C temperature, alternative base, solvent ratio) to identify the operating range.
4. Confirm the scaled reaction with duplicate runs to demonstrate reproducibility before committing to larger quantities.

The 23 kit ligands are available for individual purchase from our catalog for scale-up, with product links provided directly in the portal output. Please note that not all AI-recommended ligands from the broader 400+ database are available through our catalog, contact your account manager for sourcing guidance on ligands outside our catalog.

Does Catalexis provide guidance or checklists for robustness testing (temperature/base/solvent sweeps)?

While Catalexis does not currently provide a formal robustness testing checklist, the Phase 2 BayBE optimization engine can be configured to systematically sweep temperature, base, and solvent variables as part of your optimization campaign, effectively building a data-informed robustness picture rather than running a fixed grid screen.

Which reaction classes are currently in scope, and which are not?

Currently in scope:

1. Buchwald-Hartwig aminations (C-N bond formation)
2. Suzuki-Miyaura couplings (C-C bond formation)
3. Any Pd-catalyzed cross-coupling reaction where monophosphine ligand choice is a key variable (Phase 1 kit)
4. Any chemical reaction for which conditions can be encoded as discrete/continuous variables and an outcome metric is measurable (Phase 2 BayBE, kit-free)

Outside current scope of the physical kit:

1. Reactions requiring bidentate, NHC, or non-phosphine ligands
2. Nickel-catalyzed, copper-catalyzed, or other non-Pd metal-catalyzed cross-couplings requiring a different ligand set

Does the platform support non-phosphine ligands or non-Pd catalysis? If not, what's on the roadmap?

The physical kit is currently limited to monophosphine ligands and the Pd₂(dba)₃ metal source. Non-phosphine ligands (NHC, bidentate phosphines, etc.) and non-Pd metals (Ni, Cu, Rh) are not included in the current kit offering.

However, the BayBE optimization engine (Phase 2) is fully metal/ligand-agnostic, any catalyst or ligand can be included as a custom variable in a BayBE campaign, regardless of whether it is supplied by MilliporeSigma.

Note: Roadmap plans for expanded kit formats (e.g., NHC ligand plates, Ni-catalysis kits) should be confirmed with the Catalexis team via Catalexis@sigmaaldrich.com.

How do I benchmark the model's benefit on my chemistry?

A practical benchmarking approach:

1. Experiments saved: Count total reactions run with Catalexis to reach your target yield vs. a full grid screen of your standard ligand set.
2. Time to result: Track calendar days from first experiment to final condition selection, compared to historical projects.
3. Ligand discovery rate: Record how many top-ranked ligands from Catalexis were candidates you would not have independently selected.
4. Yield uplift: Compare the final optimized yield to the best result from your initial 23-ligand screen.

Can I run an A/B comparison (traditional vs. Catalexis-guided) to quantify impact before wider rollout?

Yes, running a parallel traditional campaign alongside a Catalexis campaign on the same substrate is a valid and recommended approach for internal validation. This allows direct comparison of:

1. Number of experiments required to achieve the same yield target
2. Diversity and novelty of ligands identified
3. Total material and time consumed

Can we export/import plate maps and results to or from our ELN or LIMS?

Results from Catalexis can be exported as CSV files, which can be imported into most ELN (Electronic Laboratory Notebook) and LIMS systems. Pre-formatted plate map templates and direct ELN integrations via API or plug-in may be available.

Is there an API or flat-file template for batch uploads from HTE instrumentation?

The availability of a programmatic API or standardized flat-file upload template for batch data ingestion from HTE liquid handlers or plate readers (e.g., HPLC, UV-Vis) should be confirmed with the Catalexis team. This is particularly relevant for automated HTE labs.

Are there recommended plate layouts for 23-ligand screens with internal standards?

The 1 µmol Screening Plates include 24 wells, 23 ligand wells plus one additional well that can be used for a control or internal standard. For manual setups, any 24-well or 96-well plate format is compatible.

Note: Guidance on plate maps, internal standard placement, and control well design can be provided upon request. Conatct Catalexis team at Catalexis@sigmaaldrich.com.

Can Catalexis be used alongside retrosynthetic planning software?

Yes, Catalexis is complementary to retrosynthetic planning tools. A typical integrated workflow:

1. Use SYNTHIA^® Retrosynthesis Software to identify optimal synthetic routes and key coupling steps.
2. Use Catalexis to rapidly screen and optimize catalyst/ligand conditions for those coupling steps.
3. Use AIDDISON^™ for earlier-stage drug discovery and molecule design.

Our Digital Chemistry platform, which includes SYNTHIA^®, AIDDISON^™, and Catalexis, is designed to support end-to-end digital reaction development.

How does Catalexis fit within Digital Chemistry solutions?

Catalexis is part of our Software & Digital Platforms portfolio, within the Digital Chemistry suite alongside:

1. SYNTHIA^®, AI-powered retrosynthetic route planning
2. AIDDISON^™, AI drug discovery (molecular design, docking, pharmacophore screening)

Together, these tools address the full synthetic chemistry workflow: route design (SYNTHIA^®) to molecule optimization (AIDDISON^™) and finally reaction condition optimization (Catalexis).

Which export formats are available for results and recommendations?

CSV export of ligand ranking results and optimization recommendations is confirmed. Data stored in your account can be accessed for future reference and export.

Can I programmatically fetch results for downstream QSAR/DFT analysis?

Results can be exported as CSV for use in external computational workflows (QSAR, DFT descriptor correlation). A programmatic API for automated result retrieval is not confirmed in current platform documentation.

Do I retain full control of my IP and data?

Yes. A key design principle of Catalexis is IP protection by design. The platform is intentionally built so that:

1. No structural information about your substrate, product, or synthetic target is required or collected.
2. You upload only numerical yield values (0-100 integers) for Part 1, and condition labels (not structures) for Part 2.
3. Your data is stored in your account and is not shared with other users or used to train models shared publicly.

Your compound identities, synthetic routes, and proprietary reaction knowledge remain entirely confidential.

Is the portal secure?

The Catalexis portal is hosted as a secure web application. Standard security practices, including HTTPS encryption for data in transit and access controls, apply.

Which browsers and versions are supported? Are there firewall or VPN requirements?

The Catalexis portal at Catalexis-app.com is a modern web application accessible from standard browsers.

No VPN is typically required. Users should ensure that access to the Catalexis website is allowed by their organization's firewall and network security settings.

Where is my data stored, and how can I request data deletion?

Note: Data storage location (region/jurisdiction), retention policies, and the process for requesting data deletion at end of project should be confirmed with the Catalexis team. This is particularly important for customers in regulated environments (GxP) or under GDPR/CCPA.

 Can we enable SSO / Okta / Azure AD for user management?

Single sign-on (SSO) integration with enterprise identity providers (Okta, Microsoft Azure AD, etc.) is not confirmed in current platform documentation. 

What exactly is saved when I upload results? Can we anonymize project names?

When you upload results, the portal stores your numerical yield/condition data and associates it with your account and project. No molecular structures are captured. You are free to use anonymized or coded project names (e.g., internal project codes) to further protect your IP.

How long are datasets retained, and can retention be configured per customer?

Dataset retention periods and whether retention can be configured on a per-customer or per-project basis should be confirmed with the Catalexis team.

Is project-level export available for archiving (CSV/XLSX/JSON)?

CSV export of project results is confirmed. Full project-level archiving (including all iterations, recommendations, and metadata) in structured formats should be confirmed with the Catalexis team.

Can multiple colleagues collaborate in one workspace with different permissions (viewer/editor/admin)?

Multi-user workspace functionality, including role-based access control (viewer/editor/admin) and team-level project sharing, should be confirmed with the Catalexis team.

Are change logs kept for uploads, edits, and token use? Can I export an audit trail for QA?

Audit trail capabilities, including logging of data uploads, edits, and token redemptions with timestamps and user attribution, are not confirmed in current documentation. This is important for GxP-adjacent workflows. 

How do I purchase Catalexis kits and portal access?

Catalexis kits are available through Aldrich Market Select (AMS). To request a quote or place an order, submit the Catalexis Request Form or contact your account manager.

Can I combine Catalexis with chemical sourcing services?

Yes. All 23 ligands in the kit and all recommended ligands are available individually through Sigma-Aldrich. The platform naturally integrates with the broader MilliporeSigma catalog:

1. Individual phosphine ligand purchases for scale-up
2. Pd catalyst sources
3. Custom synthesis services
4. Aldrich Market Select (AMS) for bulk and specialty chemical orders

How is access licensed, per user, per team, or token-based?

Portal access is token-based. Each kit includes 5 portal tokens, where 1 token = 1 ligand ranking report (one 23-ligand screen = one set of ranked recommendations). Registration on the portal (Catalexis-app.com) is free with no upfront payment. Tokens are redeemed when you submit a ligand screening dataset.

Additional tokens can be purchased separately. Custom token/credit bundles are available, and pricing is generated in line with customer needs. Request a quote via the Catalexis Request Form.

How many portal tokens come with each kit, and how are tokens redeemed? Can I buy additional tokens separately?

Each Catalexis kit (both formats) includes 5 portal tokens. Token redemption:

1.  1 token is redeemed each time you upload a 23-ligand yield dataset and generate a ranked ligand recommendation report.
2. BayBE optimization iterations (Part 2 recommendation cycles) may consume tokens separately, confirm the exact token model with your sales contact.

To place an order: submit the Catalexis Request Form or contact your account manager.

Do tokens expire or carry over between projects?

Token expiry terms, including whether tokens have a fixed validity period and whether unused tokens roll over to new projects, should be confirmed at time of purchase. Submit a query to the Catalexis Team at Catalexis@sigmaaldrich.com.

What lead times apply for kits vs. plates? Can I mix kit types in one program?

Lead times for the 50 mg Ligand Kit and 1 µmol Screening Plates, and the ability to mix kit formats within a single program order, can be confirmed by submitting the Catalexis Request Form or contacting your account manager.

Is the portal available in multiple languages and regions?

The Catalexis portal is accessible globally. The platform and the Sigma-Aldrich product page are currently available in English.

Note: Multi-language portal support and regional availability should be confirmed with the Catalexis team, at Catalexis@sigmaaldrich.com particularly for customers in China, Japan, or other non-English markets.

How often is the model updated, and will my historical projects be re-scored?

For information on model update frequency, release notes, and version history, contact Catalexis@sigmaaldrich.com.

Do you offer demos, training, or method-setup guidance?

Yes. We offer multiple ways to get started with Catalexis:

Request a live portal demonstration via the Catalexis Request Form.

Where can I find resources and literature?

The following resources are available:

1. Catalexis Portal: Catalexis-app.com
2. Product page: Catalexis Product Page
3. Publication: Design and Application of a Screening Set for Monophosphine Ligands in Cross-Coupling, RSC Digital Discovery (2025)
4. Product video: Available on the Catalexis product page

What are the standard response times for support (SLA)?

Formal support SLAs can be confirmed at time of purchase or onboarding. To discuss your support requirements, submit the Catalexis Request Form.