Reading Results

Every entry in PlasticDB comes from an experiment in which a microorganism or a protein was tested against a plastic. PlasticDB records the reported outcome of that test, not just the fact that it happened. Each test has one of three results:

• Positive — degradation of the plastic was observed.
• Inconclusive — the test was run but the outcome was ambiguous or could not be clearly interpreted.
• Negative — the organism or protein was tested and no degradation activity was detected.

Negative and inconclusive results are valuable: knowing that a candidate did not degrade a plastic helps avoid repeated experiments and sharpens the picture of what truly works. Because of this, PlasticDB is careful never to present a negative result as if it were evidence of degradation.

On a plastic page, microorganisms and proteins are split into separate tables by how they performed against that plastic:

An actor is placed in the best bucket it qualifies for: a single positive result is enough to list it as degrading, even if other papers disagree.

Sometimes different studies report different outcomes for the same pairing. The small badges next to a result tally the supporting papers on each side:

Positive 3+ 1? 2−

This example means three papers reported degradation, one was inconclusive, and two reported no activity. On microorganism and protein pages a pairing with more than one polarity is flagged Mixed so you can see the disagreement at a glance.

Headline counts across the site — the “Degrading Microorganisms” / “Degrading Proteins” numbers on plastic pages, the per-plastic counts in the plastics table, the totals on the statistics page, and the country map — count degraders only: distinct microorganisms or proteins with at least one positive result.

For a microorganism, “at least one positive result” has two readings, and the site uses the wider one. An organism reaches a plastic either because it was assayed whole, or because an enzyme purified from it was — PlasticDB v1 credited the organism in both cases, the plastic pages and the taxonomy tree do the same, and so does the landing page. The landing card therefore shows the split it depends on: of the microorganism headline, the first number is organisms assayed whole and the second is organisms credited solely through one of their enzymes, which the organism itself was never tested for. Each half links to the list that returns exactly it.

Every organism-facing surface takes an evidence_scope: organism_or_enzyme (the default, the wider reading), whole_organism, or enzyme_only. The latter two partition the first, so the two halves always add up to the whole.

Entities whose only evidence is negative or inconclusive are not included in these totals; they still appear in the relevant “inconclusive” or “tested — no activity” tables on the plastic page, and every record remains searchable.

For proteins, one question is asked everywhere. The Proteins list opens on Positive, the landing card counts the same thing, and GET /api/v1/proteins and the proteins CSV / JSON / FASTA exports default to it too — so the number on the page, the number on the home page and the number of rows in your download agree.

Nothing is hidden by that default, and everything it leaves out is one parameter away. Set Result to Everything, negatives included — ?include_negative=true — for every public record, including entities tested with no activity found and those still awaiting their first assay; or ask for a class directly with ?polarity=negative or ?polarity=inconclusive. The filter always names the question, and the page URL always records it.

Expand the Evidence section on any microorganism, protein or plastic page to see the underlying tests — one row per experiment with its result, the source paper, and a drawer of experimental conditions and observations.

Every result in PlasticDB is traceable to published literature with a DOI. Unlike databases that include computationally predicted activity, each entry here is backed by a published experiment. The Data Model page explains how these records connect, and the About page covers data provenance.