Conformal Selection 用于 CRISPR 脱靶设计的可行性检验
What Conformal Selection Can Certify in CRISPR Off-Target Design:Benchmark Leakage, Free Baselines, and a Finite-Sample Cohort Requirement
这篇论文算了笔明白账：现有公开数据只有 10 条验证向导，想做 conformal 认证得要 2 万条，根本不够。还揪出了 CRISPR 基准里的数据泄漏，简单基线就能打平训练模型。做基因编辑机器学习评测的人应该看看。
一项 arXiv 论文检验了 conformal selection 能否为机器学习驱动的 CRISPR 向导设计提供分布无关、有限样本的错误率保证。核心约束是算术层面的：conformal p 值的下限为 1/(校准队列规模+1)，因此要认证候选池，校准队列规模至少要达到候选池规模除以容忍错误率。以 2000 个候选、10% 错误率为例，需要 2 万条已验证向导，而公开 CIRCLE-seq 队列只有 10 条，导致超过 96% 的数据划分中认证集合为空。
论文还发现基准测试存在数据泄漏：标准 CIRCLE-seq 文件在 80 条 on-target 序列下列出 10 条向导，按序列列分组的交叉验证实际在同一条向导上训练和测试。修正为对级别划分后，判别力与逐向导保真度均大幅下降。
去泄漏后，免训练的位置权重打分和基因组近匹配位点计数这两个简单基线，表现不输训练出的 oracle。作者给出了修正后的评测协议、新模型必须超越的基线，以及通过认证所需的队列规模。
What Conformal Selection Can Certify in CRISPR Off-Target Design:Benchmark Leakage, Free Baselines, and a Finite-Sample Cohort Requirement
Conformal selection promises distribution-free, finite-sample control of the error rate of a decision, and has been proposed as a safety layer for machine-learning CRISPR guide design. We ask whether such a certificate is attainable from the validated off-target data that exist. The constraint is arithmetic: a conformal p-value cannot fall below one divided by the calibration cohort size plus one, so a Benjamini-Hochberg selection certifies a single candidate only when that cohort is at least the candidate pool size divided by the tolerated error rate. Public cohorts supply ten validated guides, whereas a pool of two thousand at a ten percent error rate needs twenty thousand; the accepted set is consequently empty in over ninety-six percent of splits. Two benchmarking findings explain why this went unnoticed. The standard CIRCLE-seq file lists ten guides under eighty on-target strings, so grouping cross-validation on the sequence column trains and tests on one guide; corrected, pair-level discrimination and per-guide fidelity collapse. With that leakage removed, a training-free position-weight score and a bare count of near-match genomic sites are not beaten by a trained oracle. We give the corrected protocol, the baselines a new model must beat, and the cohort size a certified pipeline needs.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：CRISPR、conformal selection、CIRCLE-seq、数据泄漏，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.LG」（@Mohammad Malekpouri, Somayeh Lotfi），经 AITOP 多维评分管线结构化提炼。
