OpenWorm 是 2011 年启动的开源计划，至今约 13 年，目标是把线虫 C. elegans 做成「分子级」数位分身。现有流程（以 c302 等框架驱动肌肉与流体环境）只能勉强呈现前进蠕动：大约需要 10 小时运算，才能生成约 5 秒行为。由于团队不做自家实验、主要靠汇整文献资料，进度高度受限于可取得且可整合的实验输入。

C. elegans 看似简单却难以「合成」：全身细胞数不到 1,000，其中只有 302 个是神经元，但仍能产生多样行为。即使此物种相关研究已带来至少 4 座诺贝尔奖，且神经连线图谱近乎完整，研究者仍缺乏能解释行为的动力学（像「操作手册」而非「电路图」）。历史反复显示落差：Brenner 团队切片与重建耗时 13 年（1986 年发表里程碑成果）；其后几乎每 10–20 年就有一次新尝试；2003 年提出的「重大挑战」甚至只做到建模外阴部；连 Cohen 投入约 20 年的模型也仍解不出倒退运动。

最新的「逆向工程」构想（37 位共同作者）主张逐一活化每个神经元，量测它对其余 301 个神经元的影响，并在平行实验中重复「数十万次」以搜集足够资料；规模预估需约 20 个实验室协作，可能长达 10 年、花费数千万美元、使用约 10 万到 20 万条真实线虫，资料量甚至可能超过过去所有 C. elegans 研究总和。支持者把它比作 NASA 式登月计划，推动自动化、巨量资料与机器学习；怀疑者则指出，最终可能只是昂贵地重现一个 1 毫米小动物此刻就能做到的事，同时把「完美模拟算不算活著」这类问题推到台前。

OpenWorm, an open-source effort launched in 2011, has spent 13 years trying to build a molecule-level digital twin of the nematode C. elegans. Its current pipeline (c302 driving muscles in fluid dynamics) can reproduce only rudimentary motion: about 10 hours of computation to generate ~5 seconds of forward squirm. The project relies on integrating published experimental data rather than running its own lab, so progress tracks the availability and standardization of measurements.

C. elegans looks simple but resists synthesis: it has fewer than 1,000 cells, yet 302 neurons yield flexible behaviors across environments. Despite at least 4 Nobel Prizes tied to the organism and a complete wiring map, researchers still lack the dynamics—an “operating manual.” Historic and recurring attempts show the gap: Brenner’s team spent 13 years (culminating in 1986); later efforts recur every 10–20 years; a 2003 “grand challenge” stalled at modeling a vulva; even after ~20 years of modeling, backward locomotion remains unresolved.

A new reverse-engineering proposal (37 coauthors) would stimulate each neuron individually and quantify effects on the other 301, repeated hundreds of thousands of times in parallel across ~20 labs. Estimates: up to 10 years, tens of millions of dollars, and roughly 100,000–200,000 worms—yielding more C. elegans data than all prior work combined. Advocates frame it as a NASA-style moonshot that advances automation and machine learning; skeptics note it may recreate, expensively, what a 1‑mm animal already does, while raising questions about what “alive” means in a perfect simulation.