Identifying opportunities for late-stage C-H alkylation with high-throughput experimentation and in silico reaction screening
Model not reportedVerification pending View verification →
Submitted summary
ORD curation of Nippa et al., DOI 10.1038/s42004-023-01047-5. Goal ord-five-20261013-jordan.
814 ORD descriptions comprise 621 main HTE records, 115 newly conducted validation experiments, 12 numeric optimization cells, 18 scale-up preparations with 26 characterized products, and 48 source-described comparison arms in 24 pairs. The 48 arms may overlap other records; 814 is not a count of unique physical syntheses. Absolute arm outcomes remain unreported and are not inferred from comparative improvement labels. All 368 synthetic decoys and 44 historical literature rows are explicitly excluded.
LCMS categories are area percentages, not yields or molar conversions. Uncharacterized product classes retain formula-targeted identifiers rather than invented regioisomer graphs. Characterized product graphs follow actual SI drawings; competing names/formulas/masses are disclosed. In particular, 38e's drawing differs from its aromatic name and HRMS information: the drawing is transcribed faithfully, while actual isolated identity remains uncertain. Other source conflicts including 7e/7q/37b and scale/solvent ratios are explicit.
Independent Astra QA checked all 736 primary SURF joins and 2,208 analytical values, input quantities/conditions, all 12 optimization cells, 18 preparations/26 product drawings and all 48 comparison arms. Review corrected initially annotation-only comparison coverage and clarified stereo evidence: three Indinavir bonds are source-drawn, two additional centers follow unambiguous source name/CAS identity confirmed with primary registries. Oxidant charge-balanced identity and the source sulfide/CAS mismatch are documented. Final review passes for complete available-evidence coverage with no open required fixes; source uncertainty is not claimed resolved.
Exact final files passed actual ord-schema 0.8.3 validation with no errors and a fresh offline ZIP rebuild byte-identically. All 37 archive manifest entries were independently hash-checked. Required tfp verify preflight passed; uploaded server hashes match the reviewed pair. Sources include main SI, reconstruction inputs/code, coverage/exclusion/comparison ledgers and validation. Optional NMR PDF is supplied by official URL/hash/verifying downloader, keeping the archive within upload limits without losing rebuild inputs.
No dataset/reaction IDs; per-reaction DOI and established public organizational curator provenance with AI assistance disclosed. Internal QA is not formal human curator approval. No upstream ORD ingestion or formal acceptance is claimed.
Submitted files 2
nippa-2023-minisci.binpb.gz154.8 KB · View file →nippa-2023-sources.zip4100.3 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 6
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: Identifying opportunities for late-stage C-H alkylation with high-throughput experimentation and in silico reaction screening
- Authors: David F. Nippa, Kenneth Atz, Alex T. Müller et al.
- Published: Communications Chemistry, 2023-11-20
- DOI: 10.1038/s42004-023-01047-5
- 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 1,104 reactions; HTE): 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
Late-stage Minisci-type C-H alkylation HTE (Roche/ETH). The data is SURF-formatted TSV, published as Nature Communications Chemistry Supplementary Data 2-8 and mirrored in the GitHub repo ETHmodlab/minisci (minisci/data/) and on Zenodo (10.5281/zenodo.8344586). Columns cover conditions (temperature, time, atmosphere, scale, concentration, mixing) and CAS, SMILES and equivalents for starting materials, reagent, catalyst, additive and up to two solvents. Outcomes are the columns product_mono, product_di and product_non, apparently relative fractions, plus a binary success label. Pitfalls: column definitions are not explicitly documented, so derive their meaning (e.g. whether outcome fractions are LCMS area ratios) from the paper's methods. Most products are not structurally characterized and are only labelled mono- or di-alkylation, so product SMILES must be generated from reactant SMILES and regiochemistry is often undetermined. Document how products are represented. The paper's decoy reactions (368) and literature reactions (45) are distinct subsets; decide explicitly whether to include them (decoys should likely be excluded, and literature reactions belong to other publications). Prefer the Nature SI copy (CC-BY) over the AGPL-licensed repo copy as the source. Cite DOI 10.1038/s42004-023-01047-5.
Known locations:
- https://www.nature.com/articles/s42004-023-01047-5: Article. The data-availability statement lists Supplementary Data 2-8 as SURF TSV files: all reactions, literature (45), experimental (691), literature-validation reactions, decoy (368), carboxylic acids, substrates (license: CC-BY 4.0 (gold OA article; SI assumed to follow article license, confirm))
- https://github.com/ETHmodlab/minisci/tree/master/minisci/data: Repo data dir: 20230830_minisci_pub_{all_data,hte,hte_new,literature,decoy}.tsv, template_acids.tsv, template_substrates.tsv. The HTE TSV has 620 data rows; columns include rxn_id, conditions, startingmat/reagent/catalyst/additive/solvent CAS+SMILES+eq, product_mono, product_di, product_non, binary (license: AGPL-3.0 (repository license, code-oriented))
- https://zenodo.org/records/8344587: Zenodo archive (concept DOI 10.5281/zenodo.8344586) of repo v1.0, minisci-v.1.0.zip 13.8 MB (not downloaded) (license: other-open (Zenodo metadata))
Pitfalls
Products are uncharacterized (mono/di labels only) and regiochemistry is ambiguous, so product identifiers are a curation decision rather than extracted fact. Outcomes are semi-quantitative fractions, not isolated yields. The decoy set and literature-sourced subsets must be separated. SURF column semantics are not documented.
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 →