Skip to content

Category

gene editing

327 papers

#artificial intelligence Preprint Jul 2026

Constitutional Midtraining: Content Presence Drives Alignment Gains

Post-training alignment is often shallow, eroding under fine-tuning. It remains untested as to whether constitutional midtraining interventions can produce durable alignment when cleanly isolated from post-training. We build a 394M-token constitutional corpus from Anthropic's Constitution and apply constitutional midtraining at 120B scale, where principled, values-based content is inserted into midtraining. A 2x2 design (curriculum ordering x deliberative reasoning) was used to produce four constitutionally midtrained conditions, plus a control, which were evaluated on self-generated and established benchmarks including alignment under pressure, value conflict resolution, blackmail, and emergent misalignment. All models were evaluated across three stages: post-midtraining, post-SFT, and post-benign fine-tuning. Constitutionally midtrained models outperformed the control on alignment generalization and durability, notably on blackmail: SFT instilled a blackmail propensity in all models, but constitutional midtraining blunted it, with the advantage surviving benign fine-tuning (-17.5pp). This durability did not extend to settings that required active resistance to in-context pressure or conflict, where the advantage attenuates after SFT. The presence of constitutional content at midtraining also mattered more than its structure, and constitutional midtraining incurred no capability cost, on average, at any stage (MMLU, ARC-Easy, piqa, GSM8K). A modest amount of constitutional content at midtraining could therefore yield broad, persistent alignment gains, offering a cheap, complementary addition to SFT-centered pipelines. Code, data, and models are available.

Desiree Cho, Cameron Tice, Bernie Hogan et al. · 0 citations
#gene editing Open access Aug 2026

Evaluating high-fidelity CRISPR-Cas nucleases in nucleosomal contexts using a quantitative framework

It is demonstrated that local nucleosome sequence and structure profoundly influence Cas nuclease accessibility and specificity, with HIFIv1 emerging as the top-performing nuclease for nucleosomal targets, while evoSpCas9 excelled in exposed contexts.

Christopher Handelmann, Erin Skeens, George P. Lisi et al. · 0 citations
#gene editing Open access Aug 2026

Polymorphism of the <i>VEGFA</i> (rs2010963) gene and low-dose exposure to airborne nickel as modification factors of programmed cell death in the Far North residents

Exposure to airborne nickel in the average daily dose, combined with the minor C-allele and CC-genotype of the VEGFA induces changes in indices of cell death and the immune profile in adults in the Far North (declining AnnexinV-FITC+7AAD– levels, growing levels of VEGF and IgE to nickel).

N. Zaitseva, N. Nikonoshina, O. V. Dolgikh et al. · 0 citations
#gene editing Open access Aug 2026

Navigating innovation and risk: evaluating European consumer perceptions of gene editing and pesticide use in watermelon production

Perceived risks associated with food production are a major concern among European consumers and strongly influence debates on sustainable agricultural practices. In particular, the adoption of precision-bred (gene-edited) crops and the use of alternative pesticide strategies raise questions about consumer acceptance and regulatory alignment. This study examines consumer perceptions of precision-bred watermelon and different pesticide strategies, including conventional pesticides, organic-certified pesticides, and zero-residue pesticides, with the aim of assessing perceived risks and implications for sustainable agriculture in Europe. A discrete choice experiment was conducted with 1,218 consumers from Spain, Germany, and the United Kingdom. The results show that between 22 and 27% of consumers across these countries reject precision-bred watermelon. Concerns related to gene editing and conventional pesticide use frequently overlap among consumers, indicating a broader sensitivity to production-related risks. However, conventional pesticides are generally perceived as riskier than precision breeding technologies. In addition, consumers show a stronger preference for zero-residue pesticides compared to organic-certified alternatives, suggesting that residue-related concerns play a key role in shaping preferences. Overall, acceptance levels of precision-bred crops are relatively similar across countries, despite differences in regulatory frameworks. The findings indicate relatively stronger consumer rejection of conventional pesticide use compared with alternative technological options, suggesting potential opportunities to expand the adoption of precision breeding technologies while addressing consumer risk perceptions and sustainability objectives. From a policy perspective, promoting zero-residue pesticide strategies and reconsidering regulatory approaches to gene editing, particularly within the European Union, may help align legislation with consumer preferences and sustainability goals. This study contributes novel empirical evidence to ongoing discussions on agricultural biotechnology, pesticide reduction, and sustainable food production in Europe.

Á. García, R. Nieto-Villegas, A. Rabadán · 0 citations
#gene editing Open access Aug 2026

Breast cancer in women of an industrial city: medico-demographic, hygienic and genetic aspects

Reducing the incidence and mortality rates of breast cancer requires comprehensive regional scientific approaches, including epidemiological, medico-demographic, hygienic, clinical, and molecular-genetic studies as the basis for integrated information for the elaboration and implementation of early diagnosis and prevention programs for oncological diseases.

N. Mikhailova, V. Kislitsyna, O. I. Baran et al. · 0 citations
#gene editing Review Aug 2026

Precision Lipid Management in the Era of Biotechnology and Artificial Intelligence: From Gene Editing to Smart Drug Delivery.

This narrative review critically evaluates emerging biotechnology- and artificial intelligence (AI)enhanced approaches for hyperlipidemia, emphasizing translational maturity, clinical applicability, and regulatory implications.

Esra Palabıyık · 0 citations
#gene editing Open access Aug 2026

The Gene Editing of Eukaryotic Translation Initiation Factor Binding Protein 3 and Its Potential Role in Rapid Cold Hardening of Two Invasive Fruit Flies

This work represents the first application of the CRISPR-Cas9 system in B. correcta, and provides insights into the molecular mechanisms underlying rapid cold hardening adaptation in B. correcta and a potential molecular marker for monitoring cold tolerance in field populations which could guide the development of novel control strategies that target the RCH adaptation pathways.

Yuning Wang, Rihui Yan, Xianwu Lin et al. · 0 citations
#gene editing Review Open access Aug 2026

Impact of the atmospheric air pollution level and xenobiotic biotransformation systems on the incidence of breast cancer

An increase in formaldehyde concentration in the air is associated with a gain in breast cancer incidence, especially at the early stages, and it is advisable to strengthen formaldehyde monitoring, implement molecular epidemiological screening of these polymorphisms and active detection of breast cancer in genetically determined risk groups.

O. N. Gulyaeva, Ivan E. Tereshkin, V. Kislitsyna et al. · 0 citations
#gene editing Aug 2026

Light- and X-ray-Activated Liposomes for Controlled Gene Editing and Drug Delivery.

This protocol presents methods for fabricating and characterising two distinct verteporfin (VP)-integrated lipid nanoparticle formulations and demonstrates knockout of up to approximately 326 slow-muscle fibres per zebrafish embryo via light activation and significant tumour growth suppression via X-ray-triggered doxorubicin release.

Yagiz A Aksoy, Biyao Yang, G. Liew et al. · 0 citations

From tech blogs

See all →
MIT News · Artificial Intelligence Aug 17, 2026

Q&A: Rethinking how innovation happens

In his latest book, Professor Eugene Fitzgerald examines the forces that turn breakthroughs into value — and why innovation resists simple formulas.