Founder's role text (READ): "High-fidelity repair of DNA double-strand breaks using a homologous template." Founder's key components (READ): BRCA1, BRCA2, PALB2, RAD51, ATM. In the recommender: No.
DRY
DRY: none located. Same three queries as node 19, plus fasting homologous recombination in vivo mouse RAD51 foci calorie restriction double strand break repair assay. No dry-fast study measures BRCA1, BRCA2, PALB2, RAD51, ATM or recombination function.
Disconfirming and absent evidence, first
- Heydari 2007 (NAR 35(22):7485-7496) contains no CR data for HR at all. READ. HR appears in that review only as a generic double-strand-break pathway, and a mention of Sir2 working with recombination-related proteins, which is not a CR experiment.
- van Galen et al. 2026 (Nutrients 18(2):246) reports no direct CR or DR effect on HR. READ. It raises the opposite-direction hypothesis: that DR may reduce replication-dependent repair such as HR by lowering proliferation. The review flags this as a hypothesis, not a result, citing Tang D, Wu J, Wang Y, Cui H, Tao Z, Lei L, Zhou Z, Tao S. "Dietary Restriction Attenuates Inflammation and Protects Mouse Skin from High-Dose Ultraviolet B Irradiation." Rejuvenation Res 2022;25:149-157. DOI 10.1089/rej.2021.0022; and Qiu R, Qiu X, Su M, Sun M, Wang Y, Wu J, Wang H, Tang D, Tao S. "Dietary Restriction Delays But Cannot Heal Irradiation-Induced Hair Graying by Preserving Hair Follicle Stem Cells in Quiescence." Rejuvenation Res 2023;26:242-252. DOI 10.1089/rej.2023.0037. Direction for HR under energy restriction is therefore contested and unresolved: no measured increase, and a published mechanistic argument for a decrease.
- Rejected reading. One human Ramadan proteomics study is annotated by its own authors to HR. It does not support an HR row. See 20.1 and the rejection note.
Graded rows
| Row | Modality | Species / system | Tier | Grade | What was actually measured | Direction |
|---|---|---|---|---|---|---|
| 20.1 | RAMADAN | Human serum, n=14 metabolic syndrome, single arm | H3 | C | AP5Z1 protein abundance. Not HR function. Not any founder-listed component. | Up |
| 20.2 | CR / DR | Mouse, skin and hair follicle | M | C | Nothing. Hypothesis only, derived from reduced proliferation. | Proposed down |
20.1 source, opened: Mindikoglu AL, Abdulsada MM, Jain A, Jalal PK, Devaraj S, Wilhelm ZR, Opekun AR, et al. (last author Jung SY). "Intermittent fasting from dawn to sunset for four consecutive weeks induces anticancer serum proteome response and improves metabolic syndrome." Scientific Reports 2020;10:18341. DOI 10.1038/s41598-020-73767-w. Design MEASURED from the opened paper: 14 adults with metabolic syndrome (8 male, 6 female, mean age 59). 29 consecutive days of dawn-to-sunset fasting, 6 May to 3 June 2019, after four weeks of ad libitum eating as baseline. Fast exceeded 14 hours daily. Eating and drinking allowed outside the window. No calorie restriction. Single-group observational pilot, no control arm. Nano UHPLC-MS/MS with label-free iBAQ. Significance threshold: paired t-test on log-transformed values, at least fourfold change and P < 0.05. The visible text does not mention multiple-testing correction. MEASURED result: AP5Z1 up, log2 6.201, about 74-fold, P = 0.005, end of week 4. The authors describe AP5Z1 as a helicase that "likely plays a role" in homologous recombination double-strand break repair.
REJECTION NOTE for 20.1, and this is load bearing. The paper that is most often cited as showing fasting upregulates DNA repair does not touch this node as the founder defines it:
- The complete set of proteins that paper names in connection with DNA repair is AP5Z1 (up ~74-fold), POLK (down, log2 -2.987, about 8-fold decrease, P = 0.026, end of week 4), and three H2B histones, HIST1H2BB, HIST1H2BD and HIST1H2BA (each up about 6-fold, P = 0.044, 0.044 and 0.020, measured one week after the fast, not during it).
- None of them is MLH1, MSH2, MSH6, PMS2, BRCA1, BRCA2, PALB2, RAD51, ATM, OGG1, MUTYH, APEX1, POLB, XRCC1, PARP1, XPA, XPC, ERCC2/3/4/5, DDB1, DDB2, PRKDC, KU70, KU80, LIG4, XRCC4, any FANC gene, MGMT, ATR, CHEK1, CHEK2, WEE1, TERT, TERC, ATRX or DAXX. Not one of the founder's 41-page components for nodes 19 to 27.
- The one genuine DNA polymerase in the list went down.
- The histone rows are the authors' own weakest: they say histone overexpression "may also be associated with" DNA repair, and their primary argument for the histones is longevity, not repair.
- Internal defect MEASURED on the opened page: the paper writes "POLK gene (POLQ) encodes for a specialized DNA polymerase." POLK and POLQ are different genes. The symbol usage is inconsistent and the truncated text does not show the authors resolving it. Any downstream use of this row inherits that ambiguity.
- No recombination assay of any kind was performed. Protein abundance in serum is not recombination capacity.
20.2 is an M row by construction. It is written as mechanism. It is not a fasting effect and may not be presented as one.
Open questions
1. HR has never been measured under any fasting modality in humans. The founder's five named components have never been measured under any fasting modality in any species, as far as this sweep could establish. 2. The only published directional argument for HR under energy restriction points down, not up, and rests on proliferation rate rather than on repair machinery. It is unresolved.
← 19. Mismatch repair (MMR) All 41 nodes 21. Base excision repair (BER) →
dryfasting.com has been published since 2014. This evidence record was compiled on 22 November 2026. Every row names the fasting modality the study actually ran, the system it was measured in and the grade. Where no source carries a DOI or PMID the row says so rather than supplying one. Nothing is sold on this site.