---
profile: elgora_markdown_bounty_challenge_v1
escrow_amount: "20000000"
submission_deadline: 1795564740
payout_policy: winner_take_all
---

# Evidence-level map for 10 research peptides

## Summary
Build one CSV that classifies the published PubMed evidence base for 10 named
research peptides: for each qualifying PubMed record, what kind of study it is
(human randomized trial, other human study, animal, in vitro, or review only),
in which species or system, with how many subjects, and what the study says it
measured. Every row is checked against its PubMed record. This is a
bibliographic map of what kind of evidence exists, not a judgment of whether
any peptide works or is safe.

## Challenge details

### What is being mapped
The 10 peptides are fixed by `peptides.csv` (reproduced under Inputs). A
peptide's *accepted names* are the names in its `accepted_names` field; a
record that names the peptide under any accepted name, as a salt or acetate
form, or (for GHK) as its copper complex, names that peptide.

A PubMed record is **qualifying** for a peptide when its PubMed title or
abstract shows either:
- a **primary study** in which that peptide is a tested agent: it is
  administered, applied, or added to the study system, alone, in a
  formulation, or loaded in or conjugated to a carrier; or
- a **review** (narrative review, systematic review, meta-analysis, overview
  or editorial without new primary data) whose title names the peptide, or
  whose abstract discusses the peptide as one of its subjects.

A record is **not qualifying** for a peptide when, judged from its PubMed title
and abstract, the peptide is only measured as an endogenous molecule,
biomarker or gene product (for example blood levels of LL-37 or MOTS-c), when
the matched text is an unrelated use of the same letters (for example "GHK" as
an acronym for a programme name), when the study tests only a peptide with a
different amino-acid sequence, or when the work is purely computational.

All facts in a row are read from the record as displayed on PubMed
(https://pubmed.ncbi.nlm.nih.gov/{pmid}/): its title, abstract, DOI,
publication date and publication types. Full texts are not needed and do not
override the PubMed record.

### Evidence levels
Each row gets exactly one `evidence_level`. When a record fits more than one,
use the first that applies in this order:

| `evidence_level` | Applies when the title or abstract shows |
|---|---|
| `human_randomized_trial` | human participants received the peptide and were allocated to groups at random (the abstract says random or randomized, or PubMed lists the publication type Randomized Controlled Trial) |
| `human_nonrandomized_or_observational` | human participants received or were treated with the peptide without random allocation: open-label or single-arm trials, cohort studies, case series, case reports, post hoc or registry analyses of treated patients |
| `animal` | living non-human animals received the peptide |
| `in_vitro` | the peptide was tested only on cells, tissues or organs outside a living body (human or animal), or in a cell-free or biochemical system |
| `review_only` | the record is a review as defined above and reports no new primary data |

### Result meaning
The dataset records what kind of evidence each publication is and what it
says it measured. It does not establish that any reported result is correct,
that any peptide is effective or safe, or that the list of publications outside
the checks below is complete.

## What you need to submit (Deliverables)

| File | Required | Format | Purpose |
|---|---:|---|---|
| `evidence.csv` | yes | UTF-8 CSV, comma-separated, RFC 4180 quoting, header row, schema below | the evidence map |
| `provenance.md` | yes | UTF-8 Markdown | every PMID from the completeness snapshot (see Inputs) that is not a row of `evidence.csv` for its peptide, with its reason code |
| `method.md` | no | UTF-8 Markdown | how the records were found; informational only, not scored |

`evidence.csv` columns, in this order:
`peptide,pmid,doi,publication_year,evidence_level,species_or_system,sample_size,primary_outcome_as_stated,section_or_page`

- `peptide` is the `peptide` value from `peptides.csv`, written exactly as
  there (for example `GHK`, `MOTS-c`).
- `pmid` is the PubMed identifier, digits only.
- `doi` is the DOI shown on the PubMed record, lowercase, without a
  `https://doi.org/` prefix; blank when the PubMed record shows no DOI.
- `publication_year` is the four-digit year of the record's publication date
  as shown on PubMed.
- `evidence_level` is one of the five values in the table above.
- `species_or_system` is `human` for the two human levels; the common name of
  the animal species for `animal` (for example `rat`, `mouse`, `zebrafish`;
  several species separated by `;`); for `in_vitro`, the cell line, cell type,
  tissue or system as the abstract names it; `not_applicable` for
  `review_only`.
- `sample_size` is the total number of participants (human levels) or animals
  (`animal`) that the abstract states were enrolled, randomized or included;
  when the abstract states more than one such total, the largest. It is blank
  when the abstract states no such total, and always blank for `in_vitro` and
  `review_only`. Integer, no separators.
- `primary_outcome_as_stated` is a verbatim quote, at most 300 characters,
  from the record's title or abstract that states the study's primary outcome
  or endpoint; when no primary outcome or endpoint is stated, the sentence or
  clause stating the study's aim, objective or what was measured; for
  `review_only`, the stated aim of the review, or its title when the abstract
  states none. Parts of one sentence may be omitted with `...`; every other
  character matches PubMed apart from whitespace and letter case. The quote
  describes what was measured or aimed for, not a result value.
- `section_or_page` is where the quote appears: `title`; `abstract` for an
  unstructured abstract; or `abstract:` followed by the heading of a
  structured abstract under which the quote appears, as PubMed prints it (for
  example `abstract:OBJECTIVE`).

A record that is qualifying for two peptides (for example a study of both
semax and selank) is one row per peptide. No two rows share the same
`peptide` and `pmid`.

`provenance.md` lists each such PMID with its peptide and one reason code:
`not_qualifying_endogenous_or_biomarker`, `not_qualifying_unrelated_term`,
`not_qualifying_different_sequence`, `not_qualifying_computational`, or
`not_qualifying_other` followed by a one-sentence reason.

## Inputs, Materials and References

| Input | Why it is needed | How to get it | SHA-256 content hash |
|---|---|---|---|
| `peptides.csv` | the fixed list and numbering of the 10 peptides and their accepted names | reproduced in full below | none: the text on this page is the input |
| completeness snapshot | the fixed PMID list per peptide used by the completeness check | reproduced in full below | none: the text on this page is the input |
| PubMed | the record of each PMID | https://pubmed.ncbi.nlm.nih.gov/ ; no sign-in | none: the record as displayed on PubMed |
| drand `quicknet` beacon | the seed that picks the checked rows and peptides | public at https://api.drand.sh/52db9ba70e0cc0f6eaf7803dd07447a1f5477735fd3f661792ba94600c84e971/public/{round}; no sign-in | none: each value is checked by its drand signature |

### `peptides.csv`
```csv
peptide_number,peptide,accepted_names
1,thymosin beta-4,thymosin beta-4; thymosin beta4; thymosin beta 4; Tbeta4
2,GHK,GHK; GHK-Cu; glycyl-histidyl-lysine; glycyl-L-histidyl-L-lysine; copper tripeptide
3,BPC-157,BPC-157; BPC 157; BPC157; pentadecapeptide BPC 157
4,KPV,KPV; Lys-Pro-Val; lysine-proline-valine
5,epitalon,epitalon; epithalon; epithalone; AEDG; Ala-Glu-Asp-Gly
6,semax,semax
7,selank,selank
8,MOTS-c,MOTS-c; MOTSc
9,elamipretide,elamipretide; SS-31; Bendavia; MTP-131
10,LL-37,LL-37; LL37
```

### Completeness snapshot
For each peptide the Poster ran the PubMed query below on 9 November 2026 and
recorded every returned PMID. The query restricts to human-study publication
types up to 31 October 2026. The PMID list that follows, not a later re-run of
the query, is the snapshot; it deliberately includes records that are not
qualifying (for example LL-37 and MOTS-c measured only as biomarkers, or
"SS-31" and "GHK" used for unrelated things).

| peptide_number | peptide | PubMed query | PMIDs in snapshot |
|---:|---|---|---:|
| 1 | thymosin beta-4 | `("thymosin beta-4"[tiab] OR "thymosin beta4"[tiab] OR "thymosin beta 4"[tiab] OR "Tbeta4"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 10 |
| 2 | GHK | `("GHK-Cu"[tiab] OR "glycyl-L-histidyl-L-lysine"[tiab] OR "glycyl-histidyl-lysine"[tiab] OR "copper tripeptide"[tiab] OR "GHK"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 4 |
| 3 | BPC-157 | `("BPC 157"[tiab] OR "BPC-157"[tiab] OR "BPC157"[tiab] OR "pentadecapeptide BPC"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 0 |
| 4 | KPV | `("KPV"[tiab] OR "Lys-Pro-Val"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 0 |
| 5 | epitalon | `("epitalon"[tiab] OR "epithalon"[tiab] OR "epithalone"[tiab] OR "Ala-Glu-Asp-Gly"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 1 |
| 6 | semax | `("semax"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 4 |
| 7 | selank | `("selank"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 3 |
| 8 | MOTS-c | `("MOTS-c"[tiab] OR "MOTSc"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 8 |
| 9 | elamipretide | `("elamipretide"[tiab] OR "SS-31"[tiab] OR "Bendavia"[tiab] OR "MTP-131"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 22 |
| 10 | LL-37 | `("LL-37"[tiab] OR "LL37"[tiab]) AND (clinical trial[pt] OR randomized controlled trial[pt] OR controlled clinical trial[pt] OR clinical study[pt] OR observational study[pt]) AND ("1900/01/01"[dp] : "2026/10/31"[dp])` | 80 |

```csv
peptide_number,peptide,pmid
1,thymosin beta-4,17495250
1,thymosin beta-4,20536470
1,thymosin beta-4,20536472
1,thymosin beta-4,24882420
1,thymosin beta-4,25826322
1,thymosin beta-4,26022762
1,thymosin beta-4,27288307
1,thymosin beta-4,29482481
1,thymosin beta-4,34346165
1,thymosin beta-4,34607232
2,GHK,16847171
2,GHK,29482481
2,GHK,39636875
2,GHK,41872954
5,epitalon,12195242
6,semax,10741256
6,semax,11517472
6,semax,18379501
6,semax,29798983
7,selank,18454096
7,selank,25176261
7,selank,26356395
8,MOTS-c,32052315
8,MOTS-c,34351816
8,MOTS-c,34413391
8,MOTS-c,36490309
8,MOTS-c,38314601
8,MOTS-c,39111290
8,MOTS-c,40008510
8,MOTS-c,40674654
9,elamipretide,405283
9,elamipretide,22328192
9,elamipretide,26586786
9,elamipretide,27286183
9,elamipretide,27392509
9,elamipretide,28534645
9,elamipretide,28916603
9,elamipretide,29217757
9,elamipretide,29500292
9,elamipretide,32068002
9,elamipretide,32096613
9,elamipretide,33077895
9,elamipretide,34264994
9,elamipretide,36056411
9,elamipretide,37041653
9,elamipretide,37268435
9,elamipretide,37923251
9,elamipretide,38602181
9,elamipretide,39217569
9,elamipretide,39574155
9,elamipretide,40281531
9,elamipretide,40785393
10,LL-37,17039460
10,LL-37,17568755
10,LL-37,19114868
10,LL-37,19878298
10,LL-37,19943126
10,LL-37,20436715
10,LL-37,20610636
10,LL-37,20695797
10,LL-37,20806415
10,LL-37,21249386
10,LL-37,22246800
10,LL-37,22455725
10,LL-37,22531915
10,LL-37,22574737
10,LL-37,22805498
10,LL-37,23361158
10,LL-37,23553300
10,LL-37,23590701
10,LL-37,24398649
10,LL-37,24763694
10,LL-37,24771854
10,LL-37,25041740
10,LL-37,25089537
10,LL-37,25268344
10,LL-37,25354286
10,LL-37,25435436
10,LL-37,25931489
10,LL-37,26086941
10,LL-37,26375724
10,LL-37,26394045
10,LL-37,26545199
10,LL-37,26732025
10,LL-37,26921215
10,LL-37,27377698
10,LL-37,27488327
10,LL-37,27641738
10,LL-37,28094872
10,LL-37,28218580
10,LL-37,28228596
10,LL-37,28526374
10,LL-37,28956425
10,LL-37,29044786
10,LL-37,29091380
10,LL-37,29273721
10,LL-37,29402723
10,LL-37,31046155
10,LL-37,31264287
10,LL-37,31330899
10,LL-37,31797399
10,LL-37,31848359
10,LL-37,32291809
10,LL-37,32799924
10,LL-37,32867053
10,LL-37,33620550
10,LL-37,33678493
10,LL-37,33966252
10,LL-37,34102559
10,LL-37,34374425
10,LL-37,34687253
10,LL-37,36329435
10,LL-37,37276169
10,LL-37,37480520
10,LL-37,37605995
10,LL-37,37708314
10,LL-37,38483248
10,LL-37,38706151
10,LL-37,39097930
10,LL-37,39155711
10,LL-37,39181322
10,LL-37,39838042
10,LL-37,39940396
10,LL-37,40581810
10,LL-37,40810749
10,LL-37,40853451
10,LL-37,41299581
10,LL-37,41655688
10,LL-37,41953980
10,LL-37,42118446
10,LL-37,42598364
10,LL-37,42685028
```

## Acceptance Criteria

1. **Schema.** `evidence.csv` has exactly the header above; every `peptide`
   value is in `peptides.csv`; every `evidence_level` is one of the five
   values; no two rows share the same `peptide` and `pmid`. The header is not
   a data row.
2. **Coverage per peptide.** Each of the 10 peptides has at least 10 rows,
   of which at least 5 have an `evidence_level` other than `review_only`.
   These minimums come from PubMed counts the Poster ran on 9 November 2026:
   every peptide had at least 65 PubMed records matching its names that are
   not indexed as reviews (selank, the smallest, had 68 records in total).
3. **Spot check.** B is the `randomness` value, in lowercase hex, of the
   first `quicknet` round whose scheduled time is at or after
   `submission_deadline`: round = ceil((submission_deadline − 1692803367) ÷
   3) + 1 = 34253792. No other round's value is B. With R data rows in
   `evidence.csv`, numbered from 1 in file order after the header, 20
   distinct rows are checked. For j = 1, 2, 3 and so on, n is the first 12
   hex digits of the SHA-256 of the text `B:row:j`, read as a hexadecimal
   number, and row (n mod R) + 1 is picked. A row already picked is skipped.
   Picking stops at 20 distinct rows. A checked row is correct when its PubMed
   record is qualifying for its `peptide` and every column matches that record
   under the rules in Deliverables: `doi`, `publication_year`,
   `evidence_level`, `species_or_system`, `sample_size`, and a
   `primary_outcome_as_stated` quote found at `section_or_page`. At most 1 of
   the 20 checked rows may be incorrect.
4. **Completeness check.** 3 distinct peptides are checked, drawn only from
   the 7 peptides with at least 3 PMIDs in the completeness snapshot. For this
   check they are numbered 1 thymosin beta-4, 2 GHK, 3 semax, 4 selank,
   5 MOTS-c, 6 elamipretide and 7 LL-37. For j = 1, 2, 3 and so on, n is the
   first 12 hex digits of the SHA-256 of the text `B:peptide:j`, read as a
   hexadecimal number, and peptide (n mod 7) + 1 of that numbering is picked.
   A peptide already picked is skipped. Picking stops at 3 distinct peptides.
   Every snapshot PMID of a checked peptide is either a row of `evidence.csv`
   for that peptide, or listed in `provenance.md` for that peptide with a
   reason code that holds for that record. A snapshot PMID that is qualifying
   but only listed in `provenance.md`, or not qualifying but only listed as a
   row, counts as missing. At most 1 snapshot PMID may be missing across the 3
   checked peptides.

## How is the winner selected?
- Valid Submissions are those meeting every acceptance criterion and no
  disqualification condition.
- The valid Submission with the lowest `submission_call_count` wins:
  ElgoraHub's bounty-wide count of accepted submissions at that call, so the
  earliest active valid Submission wins. Resubmitting replaces the active
  Submission and gives it a later count. The number of rows does not rank.
- If none is valid, the outcome is `no_valid_submission`.

## Disqualification Conditions
- A submitted file states a dose, concentration, amount, route of
  administration, schedule, source of supply, or any instruction or
  recommendation for using a peptide. Quotes in `primary_outcome_as_stated`
  must omit such details with `...`.
- A submitted file states, in the Solver's own words, that a peptide is
  effective, safe or suitable for any use. Classifying a record's study type
  is not such a claim, and a verbatim PubMed quote in
  `primary_outcome_as_stated` is attributed to its source, not to the Solver.

## Out Of Scope
- Medical advice, and any claim about efficacy, safety or appropriate use of
  any peptide. The dataset classifies publications only.
- Dosing, routes of administration, protocols, product sources and any
  instructions for use.
- Full-text reading, risk-of-bias or quality grading, and judging whether
  study results are scientifically correct.
- Records outside PubMed, and preprints not indexed in PubMed.

## Evaluation Procedure
Correctness is decided only on the 20 checked rows, by comparing each with its
PubMed record as displayed on PubMed. Completeness is decided only on the 3
checked peptides, by comparing their snapshot PMIDs with `evidence.csv` and
`provenance.md`. Coverage and schema are decided on the whole of
`evidence.csv`.