先认清“是什么”First, identify the material
黑色节点是材料本身。例子里的病毒载体有三种上游质粒;它的来源不是单一的一条线。Black nodes represent materials. The viral vector in this example has three upstream plasmids, so its origin is not a single chain.

02 / SIGNALS NOTEBOOK · MATERIAL LINEAGE02 / 实验材料追溯设计
A scientist often has only a material ID and a question: how was this made, from which batch, in which experiment? Answering it used to take about two hours of jumping between records. I connected materials, batches and experiments in one map.科学家手里常常只有一个材料编号,想知道它怎么来的、用了哪一批、进过哪个实验。过去查清这些,要在记录之间来回翻两个小时左右。我把材料、批次和实验连成了一张图。
Explore the design看看我怎么设计 ↓一个抗体往上连着 DNA 材料、细胞和培养液。科学家除了要知道“它是什么”,还要知道“怎么得到它、用了哪一批、在哪个实验里”。谱系图要接住的,就是后面这三个问题。An antibody connects back to DNA materials, cells and culture fluid. Scientists need more than an identity: how was it made, which batch was used, and in which experiment? That is the task behind the lineage map.
研究中登记的一种对象,
例如一株细胞。A registered research entity,
such as a cell line.
同一种材料的具体批次,
有各自的记录。A specific lot of that material,
with its own record.
把来源和衍生关系连起来,
看清谁从谁而来。The source and derivative links
that show what came from what.
材料工作流 · 谱系交互 · 界面设计Material workflows · Lineage interactions · Interface design








Child materials can be registered directly from an experiment table, reusing existing information. During investigation, the direction reverses: a scientist may have only a material ID and need to recover its origins, batches and related experiments.登记衍生材料时,可以直接沿用实验材料表里的信息。到了追溯阶段,方向整个倒了过来。科学家可能只有一个材料编号,要顺着它找回来源、批次和关联实验。
用户手里有材料编号,
就该能从材料开始查。When the user has a material ID,
the investigation should start there.
原始方案把材料智能文件夹作为统一入口。在材料详情里增加谱系页签,保留已有的属性和批次页面。用户进入熟悉的记录后,就能继续查关系。The proposal uses the Materials Smart Folder as a common entry point. A Lineage tab sits inside material details, alongside existing properties and lots. Researchers can continue into relationships from a familiar record.
一份材料可以做出多批。每一批,又可能进入不同的实验。只把同名材料连起来,仍然找不回那一次实验的来路。One material can have several batches, each used in different experiments. Connecting names alone does not recover the path behind a particular experiment.
黑色节点是材料本身。例子里的病毒载体有三种上游质粒;它的来源不是单一的一条线。Black nodes represent materials. The viral vector in this example has three upstream plasmids, so its origin is not a single chain.
黄色节点是具体批次。两个批次分别通向不同的细胞材料,往后又继续分批。追到材料,还不等于追到了实验。Yellow nodes represent batches. Two batches lead to different cell materials, which branch into further lots. Reaching a material is not yet reaching the experiment.
我的方案把谱系放进材料详情,同时保留批次与关联表格。图谱负责指路,具体身份和记录仍然有地方核对。My proposal places lineage inside material details while retaining lots and related tables. The graph gives direction; records provide the detail needed to check it.
看这怎样进入界面See it in the interface把对照落到具体任务上。怎么找到材料,怎样沿关系查记录,最后带走哪些信息?这几个问题,决定图谱在 Signals 里应该怎样工作。I compare the tasks around the graph. How does someone find a material, follow a relationship into a record, and take the relevant information away? Those questions shape the workflow in Signals.
生物学家熟悉沿细胞发育关系探索。黄色路径让当前分支更突出;值得借鉴的是探索方式。Biologists are familiar with exploring cell-development lineages. The yellow path emphasizes the current branch. The useful precedent is the way people explore.
它把下游分支和当前对象的属性放在一起。我借鉴了这一点,不必每次都看大图,从材料记录也能直接找到关联对象。It pairs downstream branches with the selected object’s properties. I took away the value of reaching related entities directly from a material record, without always needing a full graph.
左侧控制筛选与显示代数,下方保留选中对象、祖先和后代记录。关系探索与逐条核验各有位置。Controls on the left set filters and generations. The selected item and ancestor/descendant records remain below the graph.
材料登记连接对象与批次;关系图支持查看上下游并跳转到关联对象。Registered entities connect to batches. Entity Map supports upstream and downstream exploration and navigation to related entities.
WormWeb、LabKey、IDBS 是前期调研中的参考;Benchling 为后续补充。WormWeb, LabKey and IDBS informed the original research. Benchling is a later addition.
从材料进入。用图看关系。
需要具体信息时,仍能回到表格和记录。Start with the material. Follow the graph.
Keep tables and records within reach.
点击节点,在菜单里展开或收起谱系。选择“查看记录”,在新页打开材料详情。图谱可拖动,下方表格随分支更新。Click a node to expand or collapse its lineage. Choose View Entity to open its record in a new tab. Drag the map to explore; the table follows the visible branches.
| 材料编号Material ID | 材料类型Material Type | 材料名称Material Name | 材料/批次Asset/Batch | 谱系关联Lineage |
|---|
材料智能文件夹负责查找,详情页负责查看一份材料。谱系作为其中的页签出现,让新能力接上原来的操作习惯。The Materials Smart Folder supports discovery; the detail page holds one material. Lineage appears as a tab within that record, extending the existing navigation.
Miro 方案里,点击节点逐步展开关系,下方的祖先与后代表格同步更新。悬停看属性,通过节点菜单打开材料详情;遇到无权限记录时给出提示。把这些状态做清楚,才能让图谱里的探索接上实际操作。The Miro flow expands relationships on demand and updates the ancestor and descendant tables with the visible nodes. Hover reveals properties; the node menu opens the material record; restricted records show a notice. These states connect exploration to real actions.
图里能顺着分支看来源与去向;表格方便逐条核对,也保留导出入口。这意味着维护两种视图,但能照顾探索和核验两种任务。后台生成报告还需要额外服务,团队把 PDF 与 API 导出拆成后续需求。The map supports following origins and descendants; the table supports checking records and accessing export. Maintaining both views adds work, but serves both exploration and verification. Server-generated reports needed additional infrastructure, so the team separated PDF and API exports into follow-up work.
This later exploration revisits information hierarchy and selection between the graph and records. It is shown separately from the original project.这版后续探索继续调整信息层级,以及图谱与记录之间的选中关系,单独放在这里,方便和原始项目区分。
我梳理属性面板的需求,整理并设计了几套交互方案。Adrian 在 Miro 中参与提问和讨论,我负责方案设计与原型迭代。I organized the requirements and designed several interaction options for the properties panel. Adrian asked questions and discussed them with me in Miro; I owned the design and prototype iterations.
我推荐第三种,先筛选,再勾选。材料类型和属性放在一起,先缩小查找范围。我随后更新原型,补上“全选/全不选”,让批量调整更直接。I recommended Option 3: filter first, then select. It brought material types and attributes together while narrowing the list. I then updated the prototype with select all / select none for batch changes.
点击右上角的「Show/Hide Properties」打开面板,选择属性后保存,表格随之更新。Click Show/Hide Properties above the table to open the panel. Choose a different set of properties, then save to see the table change.
交互演示 · 示例数据Interactive demo · Example data保留选择的自由
把重复操作交给批量处理Keep the choice
Batch the repetitive work
实际产品用“全部添加 / 全部移除”(Add All / Remove All)完成批量调整,已选属性还可以排序。Add All and Remove All handle bulk changes. Selected properties can also be reordered.
产品经理和人力团队协调公司内部实验室科研人员,并与欧洲、美国客户开展线上访谈与任务验证。这个数字衡量的是材料谱系追溯任务。PM and HR coordinated internal laboratory scientists, alongside online interviews and task validation with European and US clients. This result measures the material-lineage tracing task.
净推荐值提升 5 分Net Promoter Score increased by 5 points
面向 1000 多名用户开展线上问卷,了解他们是否愿意推荐这项功能。An online survey covering over 1,000 users measured their willingness to recommend the feature.
大约五个月,我和美国的产品经理 Adrian Papas、10 位研发工程师一起,负责材料工作流、线框、交互原型、用户测试和最终界面。回头看,我最在意的一个取舍,是把那张表格留了下来。图谱好看,能让人看见关系;表格朴素,能让人逐条核对、导出带走。科学家两样都要,我就两样都留了下来。Over about five months, working with US-based PM Adrian Papas and ten engineers, I owned the material workflows, wireframes, interaction prototypes, user testing and final UI. Looking back, the trade-off I care most about is keeping the table. The map is beautiful and shows relationships. The table is plain, and it’s where scientists check each record and take data away. They needed both, so both stayed.
Signals Notebook