Phylochemical mapping of natural products onto the plant tree of life using text mining and large language models.
Bibliographic record
Abstract
Plants produce a staggering array of chemicals that are the basis for organismal function and diversity and also provide essential human nutrients and medicine. However, it is poorly defined how these compounds have evolved and are distributed across the diverse lineages of the plant kingdom, hindering a systematic view and understanding of plant chemical diversity. Recent advances in plant genome/transcriptome sequencing have provided a well-defined molecular phylogeny of plants, on which the presence of diverse natural products can be mapped to systematically determine their phylogenetic distribution. Here, we built a proof-of-concept workflow via which previously reported diverse tyrosine-derived plant natural products were mapped on to the plant tree of life. Plant chemical-species associations were mined from literature, filtered, evaluated through manual inspection of over 2,500 scientific articles, and mapped onto the plant phylogeny. The resulting 'phylochemical' map confirmed several highly lineage-specific compound class distributions, such as betalain pigments and Amaryllidaceae alkaloids. The map also highlighted several lineages enriched in dopamine-derived compounds, including the orders Caryophyllales, Liliales, and Fabales. Additionally, the application of large language models using our manually curated data as a ground truth set showed that post-mining manual processing steps can largely be automated with a low false positive rate. Our study demonstrates that a workflow combining text mining with language model-based processing can generate broader phylochemical maps, which will serve as a critical community resource to uncover key evolutionary events that underlie plant chemical diversity and enable system-level views of nature's millions of years of chemical experimentation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.010 | 0.005 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".