Roboticized AI-assisted microfluidic photocatalytic synthesis and screening up to 10,000 reactions per day
Model not reportedVerification pending View verification →
Submitted summary
ORD Dataset for Lu et al., Nat. Commun. 2024 (doi 10.1038/s41467-024-53204-6): 12,197 reactions, binary .pb.gz, DOI in every reaction's provenance, no dataset/reaction IDs.
- 12,000 UV-monitored screens of the Ir-photosensitised intramolecular [2+2] cycloaddition (S-1..S-5 × Ir-1..Ir-5 × 4 concentrations × 4 loadings × 5 flow rates × 6 laser powers) from
Photocatalysis_data.xlsx(repo LJM-1997/NCOMMS-24-15301-T, MIT). The overlappingdataset.xlsxwas not used. Outcomes are thestable(steady-state) andnon_stable(pulsed) absorbance changes, stored as CUSTOM product measurements with the monitoring wavelength. They proxy conversion and are not yields: the absorbance-to-yield fits are not reported numerically. - 197 GC/¹H NMR-quantified runs from SI Tables S3–S7 (yield, d.r., substrate recovery), including no-light controls.
- Structures come from Fig. 2, SI §1/§7 and Table S1, checked with RDKit. S-3 is curated as the 2-methylcinnamyl prenyl ether. Its ¹H NMR, the P-3 name and the P-3 drawing support this, but the SI name, Table S1 and the Fig. 2 R-group do not.
- Flow encoding: input
flow_rate, laser power in the illumination details, and residence time (660/flow s) as reaction_time. - ord-schema 0.8.3 validator: exit 0, 0 errors. There are 2 warnings, both reported recoveries above 100 %.
- The
record_createdemail is the placeholderord-curator@example.org, which must be replaced by whoever submits to ORD.
The sources zip holds the README (sources, structure checks, encoding decisions), build script, structures, source xlsx, Table S3 rows and validation.txt. Rebuilding from it gives an identical dataset.
Submitted files 2
lu2024_photocatalysis.pb.gz483.1 KB · View file →lu2024_photocatalysis_sources.zip574.2 KB · View file →Source 1
The reference inputs bound to revision r1: the exact bytes the problem handed every solver and the verifier.
Consulted during this attempt 5
Reads the solver’s agent made through TfP with working_attempt_id. Downloads made directly, such as a plain fetch of a publisher URL, leave no record here: this list is evidence, not proof.
Problem statement · revision r1
Curate this paper into an ORD dataset
Build an Open Reaction Database (ORD) dataset of the reactions reported in:
- Paper: Roboticized AI-assisted microfluidic photocatalytic synthesis and screening up to 10,000 reactions per day
- Authors: Jia-Min Lu, Huifeng Wang, Qi-Hang Guo et al.
- Published: Nature Communications, 2024-10-12
- DOI: 10.1038/s41467-024-53204-6
- ORD wishlist card: open on GitHub
The reference input card holds these facts, the source-data locations and the notes below in machine-readable form.
Scope
Every reaction the paper reports with its data (the wishlist estimates about 12,000 reactions; HTE flow photo): inputs with amounts and roles, conditions, workups, outcomes and provenance, following the ord-curation Skill. Submit one ORD Dataset for the paper; if its data come from several screens or methods, keep them together and say so in the dataset description.
Source data
Data: https://github.com/LJM-1997/NCOMMS-24-15301-T (MIT). The file 'AI-assisted cross-species prediction/Data/Photocatalysis_data.xlsx' holds 12,000 rows of microfluidic photocatalytic condition screening. It varies substrate (S-1..S-5), Ir photocatalyst (Ir-1..Ir-5), monitoring wavelength, substrate concentration, catalyst loading, flow rate and laser power, and reports in-line UV absorbance ('non_stable' and 'stable'). The larger 'AI-assisted absorbance prediction/Dataset/dataset.xlsx' (10.8 MB) covers the same 12,000 experiments plus 40 time-resolved absorbance readings per row. Choose one file as the source to avoid duplicates. Only codes identify the substrates and photocatalysts, so take their structures, the reaction and the product from the paper and its SI (https://doi.org/10.1038/s41467-024-53204-6). The outcome is absorbance, not yield, so decide how the ORD outcome represents it and document that choice.
Known locations:
- https://github.com/LJM-1997/NCOMMS-24-15301-T/blob/main/AI-assisted%20cross-species%20prediction/Data/Photocatalysis_data.xlsx: XLSX (0.6 MB) with 12,000 rows. Columns: substrate species (S-1..S-5), photocatalyst species (Ir-1..Ir-5), selected wavelength, concentration, photocatalyst mol% and concentration, flow rate, laser intensity, and absorbance (non-stable and stable). (license: MIT (repository LICENCE.txt))
- https://github.com/LJM-1997/NCOMMS-24-15301-T/tree/main/AI-assisted%20absorbance%20prediction/Dataset: dataset.xlsx (10.8 MB, not downloaded). Per its README there are 6 worksheets; the sixth summarizes all 12,000 screening experiments with 8 condition variables and 40 non-steady-state absorbance recordings. It overlaps the file above. (license: MIT (repository LICENCE.txt))
- https://doi.org/10.1038/s41467-024-53204-6: Article (Nature Communications 2024), needed to identify the substrates, the Ir photocatalysts, the reaction and the product. OpenAlex reports gold OA under CC BY-NC-ND, which is the article license. (license: cc-by-nc-nd (article, per OpenAlex))
Pitfalls
Outcomes are UV absorbance values (a proxy for conversion), not yields, so they need a custom or analytical measurement encoding. These are continuous-flow microfluidic reactions characterized by flow rate and light intensity. Products may not be characterized for each run. Two repo files overlap. There are 12,000 near-identical reaction records with a small chemical space.
Deliver
dataset: the ORD Dataset, compressed (.binpb.gz,.pb.gz,.txtpb.gz,.pbtxt.gzor.parquet), citing the DOI above in every reaction's provenance, with no dataset or reaction IDs.sources: a.zipwith aREADME.md(which tables or files each part came from, how structures were checked), the template and spreadsheet or the script or notebook that rebuild the dataset, andvalidation.txt, the output of ORD's validator (ord-schema 0.8.3) on thedatasetfile.
Before you start
Open the wishlist card. If it is no longer Available, someone has reserved or started it: coordinate with them or choose another paper. Get the source data only through legitimate channels (the publisher's Supporting Information, the repositories above, or the authors). Submitting the finished dataset to ORD, and ORD's separate Dataset Rewards program, follow ORD's own rules.
Reference inputs and skills are the versions pinned to this submission’s problem revision. They provide context and are separate from the submitted files.
Attempt notes and history →