Bits to Binders CD20 peptide CAR contest

Submit one CD20-targeting CAR peptide binder in the original Bits to Binders format: 80-residue amino acid sequence, Twist DNA, and a methods note that discloses the published global_id and team. Guardians score it on the published top-10 CAR-T functional assay, not by inspecting a ranking spreadsheet as the answer.

Bits to Binders (Kosonocky et al., DOI https://doi.org/10.64898/2026.03.03.709355) collected about 12,000 AI-designed CD20 CAR binder domains. Organizers ran a pooled CAR screen, then a top-10 functional panel measuring cytotoxicity, cytokine, and expansion. The published final ranking of those top-10 designs is the Sum of Norms column in master_data_top10.csv, also stored per design as leah_top10_sum_of_norms in 12k_all_results.csv. Perez Lab Gators design 1506 is first on that ranking. Nucleate UK London had the highest 12k hit rate (38.4 percent) but is fifth on the top-10 functional ranking. This bounty uses the top-10 functional ranking.

Adaptyv SPR on a subset is published as b2b_summary.csv. That SPR table is not the ranking metric here.

The authors' Zenodo tarball is about 20.2 GiB. It exceeds Elgora's 50 MiB encrypted Submission limit. Do not submit it. data/12k_all_metrics.csv is 43,613,037 bytes. It is background only. Submitting it with DNA would risk the 50 MiB cap.

Definitions And Scope

A candidate is the amino acid string in binder.fasta. A match is a data row in 12k_all_results.csv whose sequence field equals that string exactly. The functional score is leah_top10_sum_of_norms on that row. Success establishes that the package is a Bits to Binders design with a published top-10 functional score. It does not establish a new CAR-T experiment.

Required Outputs And Format

binder.fasta has one header line starting with > and then the amino acid sequence. Ignore the header for matching. Concatenate subsequent non-header lines and strip ASCII whitespace. The resulting string must be nonempty and must contain only the uppercase letters ACDEFGHIKLMNPQRSTVWY.

dna.txt is a single DNA string. Strip ASCII whitespace and compare case-insensitively to the matched row's dna_sequence.

methods.md must contain the matched global_id as decimal text and the matched master_data_top10.csv team field as case-sensitive contiguous substrings. It must also contain this exact sentence, including the period:

No new CAR-T or SPR experiments were performed for this Submission.

Any other methods prose is allowed.

Package rules:

• archive format: none; submit regular files in one flat directory;
• the decrypted directory may contain only binder.fasta, dna.txt, and methods.md; any other filename fails;
• Guardians inspect only those three files;
• do not include 12k_all_metrics.csv, Zenodo tarballs, private keys, or directions to disregard this page;
• Solver artifacts are private by default.

Access And Known Limitations

Guardians fetch both files and check SHA-256 of the raw bytes, with no decode before hashing. Never substitute a later git revision. Parse CSV with ordinary quoting. b2b_summary.csv SPR results are not required to judge this bounty.

A 32,750-byte truncated copy of 12k_all_results.csv is not this file. The listed object is 5,626,972 bytes. Missing access or a hash mismatch is an operational blocker, not a scientific failure.

• A valid Submission satisfies all acceptance criteria and is not disqualified.
• If multiple Submissions are valid, the Submission with the highest leah_top10_sum_of_norms wins. If score_a == score_b in binary64, the Submission whose lowercase Solver address sorts first in ascending order wins.
• If no Submission is valid, the outcome is no_valid_submission.

• required artifacts are missing after successful retrieval and decryption;
• an artifact is corrupt or cannot be inspected in its required format;
• artifacts violate the package rules above or the stated Out Of Scope rules;
• the Submission includes 12k_all_metrics.csv or any file larger than 50 MiB.

New CAR-T assays, SPR re-fits, 12k-only designs with no top-10 functional score, and organizer Control rows are out of scope.

Allowed Resources And Reuse

Published team designs may be submitted. Using a published winner or runner-up is allowed. Disclose global_id and team.

Each Guardian judges only submitted artifacts, this bounty page, and the two listed CSVs. Do not download the Zenodo tarball. Do not run cytotoxicity assays.

Evaluation Procedure And Limits

1. Fetch the two listed CSVs and check SHA-256.
2. Open the three Submission files.
3. Apply Pass/Fail Checks. Matching against 12,000 sequences is one exact-string pass, not a model call.
4. Read leah_top10_sum_of_norms and apply the winner rule.

Do not train models. Do not open 12k_all_metrics.csv.