100+ deterministic medical research engines: stats, survival, bioinformatics, primers.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
【ARRIVE 2.0 检查】 用户问动物实验符不符合 ARRIVE 2.0 返回:ARRIVE 条目核对 (真实算法引擎执行,非大模型生成)
This tool takes no arguments.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http sanyu-research https://sanyu.wentangkeji.com/api/mcp
【ARRIVE 2.0 检查】 用户问动物实验符不符合 ARRIVE 2.0 返回:ARRIVE 条目核对 (真实算法引擎执行,非大模型生成)
No arguments.
{
"type": "object",
"required": [],
"properties": {}
}【动物给药量计算】 用户要按体重算给药体积/剂量 返回:给药量与体积 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| dose_mg_per_kg* | number | 剂量(mg/kg) |
| animal_weight_g* | number | 动物体重(g) |
{
"type": "object",
"required": [
"animal_weight_g",
"dose_mg_per_kg"
],
"properties": {
"dose_mg_per_kg": {
"type": "number",
"description": "剂量(mg/kg)"
},
"animal_weight_g": {
"type": "number",
"description": "动物体重(g)"
}
}
}【脏器系数】 用户要算脏器系数(脏器重/体重) 返回:脏器系数 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| label | string | 脏器名 |
| body_weights | array | 体重 |
| organ_weights | array | 脏器重 |
{
"type": "object",
"required": [],
"properties": {
"label": {
"type": "string",
"description": "脏器名"
},
"body_weights": {
"type": "array",
"items": {
"type": "string"
},
"description": "体重"
},
"organ_weights": {
"type": "array",
"items": {
"type": "string"
},
"description": "脏器重"
}
}
}【动物随机分组】 用户要把动物随机分组(ARRIVE 要求) 返回:分组表 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| seed | integer | 种子 |
| groups | array | 分组名 |
| method | string | 随机化方法 |
| n_total* | integer | 动物总数 |
{
"type": "object",
"required": [
"n_total"
],
"properties": {
"seed": {
"type": "integer",
"description": "种子"
},
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组名"
},
"method": {
"type": "string",
"description": "随机化方法"
},
"n_total": {
"type": "integer",
"description": "动物总数"
}
}
}【动物实验样本量】 用户问动物实验要多少只、ARRIVE 要求怎么写 返回:每组动物数 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| alpha | number | α |
| power | number | 把握度 |
| groups | integer | 组数 |
| effect_size* | number | 效应量 |
{
"type": "object",
"required": [
"effect_size"
],
"properties": {
"alpha": {
"type": "number",
"description": "α"
},
"power": {
"type": "number",
"description": "把握度"
},
"groups": {
"type": "integer",
"description": "组数"
},
"effect_size": {
"type": "number",
"description": "效应量"
}
}
}【动物生存分析】 用户有动物死亡时间要做生存曲线 返回:KM 曲线与 log-rank (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| ctrl_times* | array | 对照组时间 |
| ctrl_events* | array | 对照组事件 |
| treat_times* | array | 给药组时间 |
| treat_events* | array | 给药组事件 |
{
"type": "object",
"required": [
"ctrl_times",
"ctrl_events",
"treat_times",
"treat_events"
],
"properties": {
"ctrl_times": {
"type": "array",
"items": {
"type": "string"
},
"description": "对照组时间"
},
"ctrl_events": {
"type": "array",
"items": {
"type": "string"
},
"description": "对照组事件"
},
"treat_times": {
"type": "array",
"items": {
"type": "string"
},
"description": "给药组时间"
},
"treat_events": {
"type": "array",
"items": {
"type": "string"
},
"description": "给药组事件"
}
}
}【抑瘤率 TGI】 用户要算抑瘤率 TGI 返回:TGI 与统计检验 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| ctrl_volumes* | array | 对照组体积 |
| treat_volumes* | array | 给药组体积 |
{
"type": "object",
"required": [
"ctrl_volumes",
"treat_volumes"
],
"properties": {
"ctrl_volumes": {
"type": "array",
"items": {
"type": "string"
},
"description": "对照组体积"
},
"treat_volumes": {
"type": "array",
"items": {
"type": "string"
},
"description": "给药组体积"
}
}
}【肿瘤体积计算】 用户有长宽数据要算肿瘤体积与生长曲线 返回:体积与生长曲线 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| days | array | 天数 |
| widths | array | 短径 |
| lengths | array | 长径 |
| group_name | string | 组名 |
{
"type": "object",
"required": [],
"properties": {
"days": {
"type": "array",
"items": {
"type": "string"
},
"description": "天数"
},
"widths": {
"type": "array",
"items": {
"type": "string"
},
"description": "短径"
},
"lengths": {
"type": "array",
"items": {
"type": "string"
},
"description": "长径"
},
"group_name": {
"type": "string",
"description": "组名"
}
}
}【ATAC 差异可及性】 用户有 ATAC-seq 要做差异开放区域 返回:差异 peak (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【贝叶斯自适应设计(qPCR)】 用户要做 qPCR 实验设计、数字孪生 PCR 返回:实验设计方案 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| target* | string | 目标基因/模板 |
| constraints | string | 约束条件 |
{
"type": "object",
"required": [
"target"
],
"properties": {
"target": {
"type": "string",
"description": "目标基因/模板"
},
"constraints": {
"type": "string",
"description": "约束条件"
}
}
}【生物等效性评价】 用户要做 BE 试验的 AUC/Cmax 比较 返回:90%CI 与等效性判定 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| test_auc_list* | array | 受试制剂 AUC |
| reference_auc_list* | array | 参比制剂 AUC |
{
"type": "object",
"required": [
"test_auc_list",
"reference_auc_list"
],
"properties": {
"test_auc_list": {
"type": "array",
"items": {
"type": "string"
},
"description": "受试制剂 AUC"
},
"reference_auc_list": {
"type": "array",
"items": {
"type": "string"
},
"description": "参比制剂 AUC"
}
}
}【临床↔基础转化配对】 用户要从临床发现找基础机制,或从机制找临床应用(Bridge) 返回:转化研究假说与实验方案 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| finding* | string | 临床发现/基础发现 |
| direction | string | clinical2basic 或 basic2clinical |
{
"type": "object",
"required": [
"finding"
],
"properties": {
"finding": {
"type": "string",
"description": "临床发现/基础发现"
},
"direction": {
"type": "string",
"description": "clinical2basic 或 basic2clinical"
}
}
}【校准曲线】 用户要评估预测模型的校准度、画校准曲线 返回:校准曲线与 Brier 分数 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| labels* | array | 真实标签 |
| n_bins | integer | 分箱数 |
| scores* | array | 预测概率(0~1) |
{
"type": "object",
"required": [
"labels",
"scores"
],
"properties": {
"labels": {
"type": "array",
"items": {
"type": "string"
},
"description": "真实标签"
},
"n_bins": {
"type": "integer",
"description": "分箱数"
},
"scores": {
"type": "array",
"items": {
"type": "string"
},
"description": "预测概率(0~1)"
}
}
}【细胞实验通用计算】 用户要做 CCK8/划痕/Transwell/克隆形成等细胞实验计算 返回:计算结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| params | object | 该工具所需参数 |
| tool_id* | string | 工具 id,如 cck8/wound-healing/transwell/colony |
{
"type": "object",
"required": [
"tool_id"
],
"properties": {
"params": {
"type": "object",
"description": "该工具所需参数"
},
"tool_id": {
"type": "string",
"description": "工具 id,如 cck8/wound-healing/transwell/colony"
}
}
}【报告规范清单核对】 用户问论文符不符合 CONSORT/STROBE 等报告规范 返回:逐条核对结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| text* | string | 论文正文 |
| design_type | string | 研究设计类型 |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "论文正文"
},
"design_type": {
"type": "string",
"description": "研究设计类型"
}
}
}【ChIP-seq Peak calling】 用户有 ChIP-seq 要做 peak calling 返回:peak 清单 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【引文格式化】 用户要把参考文献转成 Vancouver/APA/GB-T 7714 格式 返回:格式化后的引文 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| style | string | vancouver/apa/mla/chicago/gbt7714 |
| references* | array | 参考文献列表 |
{
"type": "object",
"required": [
"references"
],
"properties": {
"style": {
"type": "string",
"description": "vancouver/apa/mla/chicago/gbt7714"
},
"references": {
"type": "array",
"items": {
"type": "string"
},
"description": "参考文献列表"
}
}
}【分类性能报告】 用户要看准确率/敏感度/特异度/PPV/NPV 全套指标 返回:分类指标全套 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| labels* | array | 真实标签 |
| threshold | number | 判定阈值 |
| predictions* | array | 预测结果(0/1) |
{
"type": "object",
"required": [
"labels",
"predictions"
],
"properties": {
"labels": {
"type": "array",
"items": {
"type": "string"
},
"description": "真实标签"
},
"threshold": {
"type": "number",
"description": "判定阈值"
},
"predictions": {
"type": "array",
"items": {
"type": "string"
},
"description": "预测结果(0/1)"
}
}
}【克隆方案设计】 用户要做分子克隆、构建质粒、设计同源臂 返回:克隆方案与引物 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| method | string | Gibson / Golden Gate / 酶切连接 |
| insert_seq* | string | 插入片段序列(≥40bp) |
| vector_seq* | string | 载体序列 |
{
"type": "object",
"required": [
"insert_seq",
"vector_seq"
],
"properties": {
"method": {
"type": "string",
"description": "Gibson / Golden Gate / 酶切连接"
},
"insert_seq": {
"type": "string",
"description": "插入片段序列(≥40bp)"
},
"vector_seq": {
"type": "string",
"description": "载体序列"
}
}
}【CONSORT 流程图】 用户有筛选/随机化/失访人数,要画 CONSORT 流程图 返回:CONSORT 流程图 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| title | string | 研究题目 |
| screened | integer | 筛选人数 |
| randomized | integer | 随机化人数 |
| allocated_control | integer | 对照组分配数 |
| excluded_screening | integer | 筛除人数 |
| allocated_intervention | integer | 干预组分配数 |
| lost_to_followup_control | integer | 对照组失访数 |
| lost_to_followup_intervention | integer | 干预组失访数 |
{
"type": "object",
"required": [],
"properties": {
"title": {
"type": "string",
"description": "研究题目"
},
"screened": {
"type": "integer",
"description": "筛选人数"
},
"randomized": {
"type": "integer",
"description": "随机化人数"
},
"allocated_control": {
"type": "integer",
"description": "对照组分配数"
},
"excluded_screening": {
"type": "integer",
"description": "筛除人数"
},
"allocated_intervention": {
"type": "integer",
"description": "干预组分配数"
},
"lost_to_followup_control": {
"type": "integer",
"description": "对照组失访数"
},
"lost_to_followup_intervention": {
"type": "integer",
"description": "干预组失访数"
}
}
}【CRF 数据校验(SDTM)】 用户问 CRF 数据合不合规、能不能导出 CDISC 返回:校验结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| form_id | integer | 表单 id |
{
"type": "object",
"required": [],
"properties": {
"form_id": {
"type": "integer",
"description": "表单 id"
}
}
}【数据快速画像】 用户想先看看这份数据长什么样、分布、缺失 返回:描述统计与质量概览 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| file_id* | integer | 数据文件 id |
| study_design | string | 研究设计 |
{
"type": "object",
"required": [
"file_id"
],
"properties": {
"file_id": {
"type": "integer",
"description": "数据文件 id"
},
"study_design": {
"type": "string",
"description": "研究设计"
}
}
}【药物相互作用检查】 用户问两个药能不能一起用、有没有相互作用 返回:相互作用等级与机制 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| drug_a* | string | 药物 A |
| drug_b* | string | 药物 B |
{
"type": "object",
"required": [
"drug_a",
"drug_b"
],
"properties": {
"drug_a": {
"type": "string",
"description": "药物 A"
},
"drug_b": {
"type": "string",
"description": "药物 B"
}
}
}【证据追问(多轮)】 用户要对已录入的证据继续追问、做因果判断 返回:追问答复 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| round | integer | 轮次 |
| title* | string | 标题 |
| abstract | string | 摘要 |
| question* | string | 追问问题 |
{
"type": "object",
"required": [
"title",
"question"
],
"properties": {
"round": {
"type": "integer",
"description": "轮次"
},
"title": {
"type": "string",
"description": "标题"
},
"abstract": {
"type": "string",
"description": "摘要"
},
"question": {
"type": "string",
"description": "追问问题"
}
}
}【文献证据分级录入】 用户给一篇文献要按证据等级结构化拆解(Decision OS) 返回:结构化证据卡 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| doi | string | DOI |
| title* | string | 标题 |
| abstract | string | 摘要 |
| literature_type | string | 文献类型 |
{
"type": "object",
"required": [
"title"
],
"properties": {
"doi": {
"type": "string",
"description": "DOI"
},
"title": {
"type": "string",
"description": "标题"
},
"abstract": {
"type": "string",
"description": "摘要"
},
"literature_type": {
"type": "string",
"description": "文献类型"
}
}
}【证据报告叙述】 用户要把证据卡生成可读报告 返回:证据报告正文 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| lang | string | 可选。报告语言:zh(默认,中文)或 en(英文)。其他值按 zh 处理。 |
| result* | object | 必填。上一次 act_decision_intake / act_decision_followup 返回的 result 对象,原样传入即可(内含 causality / drug_stage / evidence_chain 等字段)。实测缺此字段返回:缺少决策结果 result |
{
"type": "object",
"required": [
"result"
],
"properties": {
"lang": {
"enum": [
"zh",
"en"
],
"type": "string",
"description": "可选。报告语言:zh(默认,中文)或 en(英文)。其他值按 zh 处理。"
},
"result": {
"type": "object",
"description": "必填。上一次 act_decision_intake / act_decision_followup 返回的 result 对象,原样传入即可(内含 causality / drug_stage / evidence_chain 等字段)。实测缺此字段返回:缺少决策结果 result"
}
}
}【差异表达分析】 用户有表达矩阵要做差异基因分析、火山图 返回:差异基因清单 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| group_a | string | A 组样本表达(CSV) |
| group_b | string | B 组样本表达(CSV) |
| gene_names | array | 基因名列表 |
| pval_threshold | number | P 阈值 |
| log2fc_threshold | number | log2FC 阈值 |
{
"type": "object",
"required": [],
"properties": {
"group_a": {
"type": "string",
"description": "A 组样本表达(CSV)"
},
"group_b": {
"type": "string",
"description": "B 组样本表达(CSV)"
},
"gene_names": {
"type": "array",
"items": {
"type": "string"
},
"description": "基因名列表"
},
"pval_threshold": {
"type": "number",
"description": "P 阈值"
},
"log2fc_threshold": {
"type": "number",
"description": "log2FC 阈值"
}
}
}【两条 ROC 的 AUC 比较(DeLong)】 用户有两个指标的 AUC,要比较哪个更好、差异有没有统计学意义 返回:两 AUC 及差值 P 值 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| labels* | array | 真实标签 0/1 |
| scores1* | array | 指标 1 的分数 |
| scores2* | array | 指标 2 的分数 |
{
"type": "object",
"required": [
"labels",
"scores1",
"scores2"
],
"properties": {
"labels": {
"type": "array",
"items": {
"type": "string"
},
"description": "真实标签 0/1"
},
"scores1": {
"type": "array",
"items": {
"type": "string"
},
"description": "指标 1 的分数"
},
"scores2": {
"type": "array",
"items": {
"type": "string"
},
"description": "指标 2 的分数"
}
}
}【取某一步的上下文】 要基于前序步骤(PICO / 设计类型 / 研究对象)起草某一步内容时,先取上下文,避免重复问用户已经填过的信息 返回:该步骤可用的前序上下文(pico / design_type / population / topic) (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| step* | integer | 要看哪一步的上下文(数据编号 1-11) |
| project_id* | integer | 课题 id |
{
"type": "object",
"required": [
"project_id",
"step"
],
"properties": {
"step": {
"type": "integer",
"description": "要看哪一步的上下文(数据编号 1-11)"
},
"project_id": {
"type": "integer",
"description": "课题 id"
}
}
}【方案设计闸门规则】 用户问「为什么这步不让我过」「方案设计要满足什么条件」「哪些是硬性要求」,或助手要解释某一步被拦的原因时先读它,确保解释与平台规则一致(不靠自己编) 返回:每步是否有硬闸门、三道跨步闸门 G1/G2/G3 的原文规则 (真实算法引擎执行,非大模型生成)
No arguments.
{
"type": "object",
"required": [],
"properties": {}
}【研究设计向导】 用户不知道自己的问题该用什么研究设计、要选方案 返回:推荐研究设计与理由 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| pico* | string | 研究问题/PICO 描述 |
| feasibility | string | 可行性条件(病例量、周期、经费) |
{
"type": "object",
"required": [
"pico"
],
"properties": {
"pico": {
"type": "string",
"description": "研究问题/PICO 描述"
},
"feasibility": {
"type": "string",
"description": "可行性条件(病例量、周期、经费)"
}
}
}【样本量计算(设计口径)】 用户在设计阶段要算样本量、问要做多少例 返回:每组例数与合计 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sd | number | 标准差 |
| alpha | number | 显著性水平,默认 0.05 |
| power | number | 把握度,默认 0.8 |
| effect | number | 预期效应量/差值 |
| dropout | number | 脱落率 |
| sample_type | string | 计算类型:两组均数/两组率/相关/生存 |
{
"type": "object",
"required": [],
"properties": {
"sd": {
"type": "number",
"description": "标准差"
},
"alpha": {
"type": "number",
"description": "显著性水平,默认 0.05"
},
"power": {
"type": "number",
"description": "把握度,默认 0.8"
},
"effect": {
"type": "number",
"description": "预期效应量/差值"
},
"dropout": {
"type": "number",
"description": "脱落率"
},
"sample_type": {
"type": "string",
"description": "计算类型:两组均数/两组率/相关/生存"
}
}
}【分布拟合检验】 用户问数据是不是正态、服从什么分布 返回:拟合优度 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data* | array | 一列数值 |
| dist | string | 分布类型 |
{
"type": "object",
"required": [
"data"
],
"properties": {
"data": {
"type": "array",
"items": {
"type": "string"
},
"description": "一列数值"
},
"dist": {
"type": "string",
"description": "分布类型"
}
}
}【种属剂量换算】 用户要把动物剂量换算成人体剂量 返回:换算后剂量与依据 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| source_dose* | number | 来源剂量 |
| source_species* | string | 来源物种 |
| target_species* | string | 目标物种 |
{
"type": "object",
"required": [
"source_species",
"target_species",
"source_dose"
],
"properties": {
"source_dose": {
"type": "number",
"description": "来源剂量"
},
"source_species": {
"type": "string",
"description": "来源物种"
},
"target_species": {
"type": "string",
"description": "目标物种"
}
}
}【多组非参数两两比较(Dunn)】 用户做了 Kruskal-Wallis 后要多组两两比较 返回:两两比较 P 值 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| groups* | object | {组名: [数值...]} |
{
"type": "object",
"required": [
"groups"
],
"properties": {
"groups": {
"type": "object",
"description": "{组名: [数值...]}"
}
}
}【通路富集(GO/KEGG)】 用户有一批基因要做 GO/KEGG 富集、找通路 返回:富集通路与 P 值 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| db_filter | string | 数据库:GO_BP/GO_CC/GO_MF/KEGG |
| gene_list* | array | 基因列表 |
| project_id | integer | 项目 id |
{
"type": "object",
"required": [
"gene_list"
],
"properties": {
"db_filter": {
"type": "string",
"description": "数据库:GO_BP/GO_CC/GO_MF/KEGG"
},
"gene_list": {
"type": "array",
"items": {
"type": "string"
},
"description": "基因列表"
},
"project_id": {
"type": "integer",
"description": "项目 id"
}
}
}【流式细胞聚类】 用户要对流式数据做无监督聚类分群 返回:细胞群与比例 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| events | array | 事件数据 |
| grid_size | integer | 网格大小 |
| n_clusters | integer | 聚类数 |
{
"type": "object",
"required": [],
"properties": {
"events": {
"type": "array",
"items": {
"type": "string"
},
"description": "事件数据"
},
"grid_size": {
"type": "integer",
"description": "网格大小"
},
"n_clusters": {
"type": "integer",
"description": "聚类数"
}
}
}【流式门控分析】 用户有流式数据要做圈门 返回:门控比例 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| events | array | 事件数据 |
| method | string | 门控方法 |
| params | object | 门控参数 |
{
"type": "object",
"required": [],
"properties": {
"events": {
"type": "array",
"items": {
"type": "string"
},
"description": "事件数据"
},
"method": {
"type": "string",
"description": "门控方法"
},
"params": {
"type": "object",
"description": "门控参数"
}
}
}【纤维化分期】 用户要做肝纤维化分期(Ishak/METAVIR 等) 返回:分期结论与解读 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| stage | integer | 分期数 |
| method | string | 分期体系:ishak/metavir/nas |
{
"type": "object",
"required": [],
"properties": {
"stage": {
"type": "integer",
"description": "分期数"
},
"method": {
"type": "string",
"description": "分期体系:ishak/metavir/nas"
}
}
}【图注生成】 用户有图要写图注、Figure legend 返回:图注文本 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| context | string | 图的背景与数据 |
| chart_type* | string | 图类型 |
{
"type": "object",
"required": [
"chart_type"
],
"properties": {
"context": {
"type": "string",
"description": "图的背景与数据"
},
"chart_type": {
"type": "string",
"description": "图类型"
}
}
}【读取数据列名】 用户要知道某个数据文件有哪些变量列(做 Table One / 建模前必看) 返回:列名与类型 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| file_id* | integer | 数据文件 id |
{
"type": "object",
"required": [
"file_id"
],
"properties": {
"file_id": {
"type": "integer",
"description": "数据文件 id"
}
}
}【森林图】 用户有 HR/OR/RR 和置信区间,要画森林图 返回:森林图(矢量) (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data* | array | 每项 {name, estimate, ci_lower, ci_upper} |
| title | string | 图标题 |
| x_label | string | X 轴标签 |
{
"type": "object",
"required": [
"data"
],
"properties": {
"data": {
"type": "array",
"items": {
"type": "string"
},
"description": "每项 {name, estimate, ci_lower, ci_upper}"
},
"title": {
"type": "string",
"description": "图标题"
},
"x_label": {
"type": "string",
"description": "X 轴标签"
}
}
}【基因-通路-疾病网络】 用户要从一批基因找通路、疾病、药物关联 返回:通路/疾病/药物关联 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| genes* | array | 基因列表 |
{
"type": "object",
"required": [
"genes"
],
"properties": {
"genes": {
"type": "array",
"items": {
"type": "string"
},
"description": "基因列表"
}
}
}【基因信息检索】 用户问某个基因是干什么的、别名、功能 返回:基因信息 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| query* | string | 基因名/关键词 |
{
"type": "object",
"required": [
"query"
],
"properties": {
"query": {
"type": "string",
"description": "基因名/关键词"
}
}
}【GEO 数据全流程】 用户给 GEO 编号要做差异分析、富集、免疫浸润全流程 返回:差异基因、富集结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| geo_id* | string | GEO 数据集编号,如 GSE12345 |
| species | string | 物种:human/mouse |
| pval_cutoff | number | P 阈值 |
| logfc_cutoff | number | logFC 阈值 |
{
"type": "object",
"required": [
"geo_id"
],
"properties": {
"geo_id": {
"type": "string",
"description": "GEO 数据集编号,如 GSE12345"
},
"species": {
"type": "string",
"description": "物种:human/mouse"
},
"pval_cutoff": {
"type": "number",
"description": "P 阈值"
},
"logfc_cutoff": {
"type": "number",
"description": "logFC 阈值"
}
}
}【Gleason 分级分组】 用户问前列腺癌 Gleason 评分对应哪个 Grade Group 返回:Grade Group 与风险分层 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| primary_grade* | integer | 主要分级 1-5 |
| secondary_grade* | integer | 次要分级 1-5 |
{
"type": "object",
"required": [
"primary_grade",
"secondary_grade"
],
"properties": {
"primary_grade": {
"type": "integer",
"description": "主要分级 1-5"
},
"secondary_grade": {
"type": "integer",
"description": "次要分级 1-5"
}
}
}【IC50 曲线拟合】 用户有药物浓度和抑制率要算 IC50 返回:IC50 值与 95%CI (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| model | string | 拟合模型:4PL/3PL |
| direction | string | inhibition 或 activation |
| responses* | array | 响应/抑制率 |
| concentrations* | array | 浓度序列 |
{
"type": "object",
"required": [
"concentrations",
"responses"
],
"properties": {
"model": {
"type": "string",
"description": "拟合模型:4PL/3PL"
},
"direction": {
"type": "string",
"description": "inhibition 或 activation"
},
"responses": {
"type": "array",
"items": {
"type": "string"
},
"description": "响应/抑制率"
},
"concentrations": {
"type": "array",
"items": {
"type": "string"
},
"description": "浓度序列"
}
}
}【IHC 免疫组化评分】 用户有 IHC 染色结果要算 H-score/Allred/阳性率 返回:H-score / Allred / 阳性率 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| stain_type* | string | 染色类型 |
| intensity_0_pct | number | 0 级占比(%) |
| intensity_1_pct | number | 1+ 占比(%) |
| intensity_2_pct | number | 2+ 占比(%) |
| intensity_3_pct | number | 3+ 占比(%) |
{
"type": "object",
"required": [
"stain_type"
],
"properties": {
"stain_type": {
"type": "string",
"description": "染色类型"
},
"intensity_0_pct": {
"type": "number",
"description": "0 级占比(%)"
},
"intensity_1_pct": {
"type": "number",
"description": "1+ 占比(%)"
},
"intensity_2_pct": {
"type": "number",
"description": "2+ 占比(%)"
},
"intensity_3_pct": {
"type": "number",
"description": "3+ 占比(%)"
}
}
}【IPTW 逆概率加权】 用户要做倾向评分加权分析 返回:权重与加权后效应 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| treatment* | array | 处理分组 0/1 |
| propensity_scores* | array | 倾向评分 |
{
"type": "object",
"required": [
"propensity_scores",
"treatment"
],
"properties": {
"treatment": {
"type": "array",
"items": {
"type": "string"
},
"description": "处理分组 0/1"
},
"propensity_scores": {
"type": "array",
"items": {
"type": "string"
},
"description": "倾向评分"
}
}
}【智能选刊 V3】 用户要按题目摘要推荐期刊、冲/稳/保三档 返回:候选期刊与三档策略 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| title* | string | 论文标题 |
| design | string | 研究设计 |
| abstract | string | 摘要 |
| keywords | array | 关键词 |
| sample_size | integer | 样本量 |
| target_tier | string | 目标档次 |
{
"type": "object",
"required": [
"title"
],
"properties": {
"title": {
"type": "string",
"description": "论文标题"
},
"design": {
"type": "string",
"description": "研究设计"
},
"abstract": {
"type": "string",
"description": "摘要"
},
"keywords": {
"type": "array",
"items": {
"type": "string"
},
"description": "关键词"
},
"sample_size": {
"type": "integer",
"description": "样本量"
},
"target_tier": {
"type": "string",
"description": "目标档次"
}
}
}【Kaplan-Meier 曲线】 用户只要 KM 曲线和中位生存,不做多因素 返回:KM 曲线与中位生存 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| times* | array | 随访时间 |
| events* | array | 事件 0/1 |
| groups | array | 分组 |
{
"type": "object",
"required": [
"times",
"events"
],
"properties": {
"times": {
"type": "array",
"items": {
"type": "string"
},
"description": "随访时间"
},
"events": {
"type": "array",
"items": {
"type": "string"
},
"description": "事件 0/1"
},
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组"
}
}
}【文献检索(站内 153 万篇 PMC)】 用户要找文献、找证据、看研究现状 返回:文献列表(含 PMCID) (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| q* | string | 检索词(英文效果更好) |
| limit | integer | 返回条数 |
{
"type": "object",
"required": [
"q"
],
"properties": {
"q": {
"type": "string",
"description": "检索词(英文效果更好)"
},
"limit": {
"type": "integer",
"description": "返回条数"
}
}
}【引文真实性核验】 用户担心参考文献是编的、要核验引用真伪 返回:逐条核验结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| text* | string | 含引用的正文 |
| topic | string | 主题 |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "含引用的正文"
},
"topic": {
"type": "string",
"description": "主题"
}
}
}【文字重合度自查】 用户问这段文字和我自己以前的文章重不重(查重) 返回:重合度与相似文献 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| text* | string | 待查文本 |
| top_n | integer | 返回条数 |
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string",
"description": "待查文本"
},
"top_n": {
"type": "integer",
"description": "返回条数"
}
}
}【线性混合模型(纵向/重复测量,公式模式)】 同一受试者在多个时间点被反复测量(治疗前后各测一次、多次随访、多个时间点重复观察、每人测三四次、每例随访若干次),要比较各组随时间的变化趋势、要看组与时间的交互;也是重复测量方差分析的现代替代(允许缺失、不必球形检验) 返回:固定效应系数/SE/Wald z/P/95%CI(按真实变量名,不是 X1/X2)、各项整体检验 omnibus、交互项检验、primary_test、ICC、方差成分、n_subjects/n_obs、组间例数、缺失与收敛告警 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data* | array | 长格式数据:每行一次观测 [{受试者ID, 分组, 时点, 结局}, ...];也接受列式 {列名: [值, ...]}。注意必须是**长表**(每行一次测量),不是每人一行多列 |
| reml | boolean | 默认 true(REML) |
| formula* | string | R 风格公式,如 '收缩压 ~ 分组 * 时点'(组×时间交互);左端=结局 |
| subject* | string | 受试者/病案号列名(随机截距的分组依据) |
| reference | object | 参照水平,如 {"分组": "对照组"} —— 报告表里「与对照组相比」必须是它 |
| re_formula | string | 随机效应,默认 '1'(随机截距);随机斜率用 '1 + 时点' |
| categorical | array | 显式声明哪些列按分类处理(如 ["分组"]);数值编码的 0/1 分组必须声明,否则会被当连续变量 |
| time_values | object | 把字符串时点映射成数值以估线性趋势,如 {"时点": {"基线":0,"4周":1,"12周":2}} |
【Meta 发表偏倚(Egger)】 用户做 Meta 分析要检验发表偏倚、画漏斗图 返回:Egger 检验 P 与截距 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| log_or* | array | 各研究 logOR |
| se_log_or* | array | 各研究 SE |
{
"type": "object",
"required": [
"log_or",
"se_log_or"
],
"properties": {
"log_or": {
"type": "array",
"items": {
"type": "string"
},
"description": "各研究 logOR"
},
"se_log_or": {
"type": "array",
"items": {
"type": "string"
},
"description": "各研究 SE"
}
}
}【代谢物通路富集】 用户有一批差异代谢物要找通路 返回:代谢通路富集 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| metabolites* | array | 代谢物列表 |
| fdr_threshold | number | FDR 阈值 |
{
"type": "object",
"required": [
"metabolites"
],
"properties": {
"metabolites": {
"type": "array",
"items": {
"type": "string"
},
"description": "代谢物列表"
},
"fdr_threshold": {
"type": "number",
"description": "FDR 阈值"
}
}
}【代谢组热图】 用户要画代谢物聚类热图 返回:聚类热图 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【代谢组 PCA】 用户要看样本分组趋势、有没有离群 返回:PCA 得分与解释度 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| groups | array | 分组标签 |
| data_json* | object | 代谢物矩阵 |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组标签"
},
"data_json": {
"type": "object",
"description": "代谢物矩阵"
}
}
}【代谢组 PLS-DA】 用户要做有监督判别分析、找差异代谢物 返回:PLS-DA 模型与 VIP (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| groups | array | 分组标签 |
| data_json* | object | 代谢物矩阵 |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组标签"
},
"data_json": {
"type": "object",
"description": "代谢物矩阵"
}
}
}【代谢组火山图】 用户要画代谢物火山图 返回:火山图 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【多重免疫荧光表型】 用户有多少免疫荧光标记要定细胞表型 返回:细胞亚群表型与比例 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| cd3_pos | number | CD3+ 数 |
| cd8_pos | number | CD8+ 数 |
| foxp3_pos | number | FoxP3+ 数 |
| pd_l1_pos | number | PD-L1+ 数 |
{
"type": "object",
"required": [],
"properties": {
"cd3_pos": {
"type": "number",
"description": "CD3+ 数"
},
"cd8_pos": {
"type": "number",
"description": "CD8+ 数"
},
"foxp3_pos": {
"type": "number",
"description": "FoxP3+ 数"
},
"pd_l1_pos": {
"type": "number",
"description": "PD-L1+ 数"
}
}
}【列线图 Nomogram】 用户要把 Logistic/Cox 模型做成列线图 返回:列线图 SVG (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| csv_data | string | CSV 文本 |
{
"type": "object",
"required": [],
"properties": {
"csv_data": {
"type": "string",
"description": "CSV 文本"
}
}
}【数据解析与列类型识别】 用户贴了一段 CSV/表格,要先看数据长什么样、每列是什么类型 返回:行列数、列类型、缺失情况 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| csv_data | string | CSV 文本 |
{
"type": "object",
"required": [],
"properties": {
"csv_data": {
"type": "string",
"description": "CSV 文本"
}
}
}【模型稳定性 PSI】 用户要评估模型换数据后稳不稳(PSI) 返回:PSI 值与稳定性判断 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| n_bins | integer | 分箱数 |
| expected_probs* | array | 开发集概率 |
| observed_probs* | array | 验证集概率 |
{
"type": "object",
"required": [
"expected_probs",
"observed_probs"
],
"properties": {
"n_bins": {
"type": "integer",
"description": "分箱数"
},
"expected_probs": {
"type": "array",
"items": {
"type": "string"
},
"description": "开发集概率"
},
"observed_probs": {
"type": "array",
"items": {
"type": "string"
},
"description": "验证集概率"
}
}
}【SHAP 单样本解释】 用户要解释某个病人为什么被判高风险 返回:SHAP 贡献图 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| csv_data | string | CSV 文本 |
| model_type | string | 模型类型 |
| sample_index | integer | 样本序号 |
{
"type": "object",
"required": [],
"properties": {
"csv_data": {
"type": "string",
"description": "CSV 文本"
},
"model_type": {
"type": "string",
"description": "模型类型"
},
"sample_index": {
"type": "integer",
"description": "样本序号"
}
}
}【NCA 非房室药代分析】 用户有血药浓度-时间数据要算 AUC/Cmax/T1/2 返回:AUC/Cmax/Tmax/T1/2/CL/Vd (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| dose | number | 给药剂量 |
| dose_unit | string | 剂量单位,如 mg/kg |
| time_points* | array | 时间点 |
| concentrations* | array | 浓度 |
{
"type": "object",
"required": [
"time_points",
"concentrations"
],
"properties": {
"dose": {
"type": "number",
"description": "给药剂量"
},
"dose_unit": {
"type": "string",
"description": "剂量单位,如 mg/kg"
},
"time_points": {
"type": "array",
"items": {
"type": "string"
},
"description": "时间点"
},
"concentrations": {
"type": "array",
"items": {
"type": "string"
},
"description": "浓度"
}
}
}【细胞核分割与计数】 用户要算切片里的细胞密度/阳性率 返回:分割统计结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| stain_type | string | 染色类型 |
| cell_density | number | 细胞密度 |
| positive_rate | number | 阳性率 |
{
"type": "object",
"required": [],
"properties": {
"stain_type": {
"type": "string",
"description": "染色类型"
},
"cell_density": {
"type": "number",
"description": "细胞密度"
},
"positive_rate": {
"type": "number",
"description": "阳性率"
}
}
}【多组学整合】 用户要把 mRNA/蛋白/代谢物一起整合分析 返回:多组学关联结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| mrna | array | mRNA 列表 |
| protein | array | 蛋白列表 |
| metabolite | array | 代谢物列表 |
| pathway_map | string | 通路映射方式 |
{
"type": "object",
"required": [],
"properties": {
"mrna": {
"type": "array",
"items": {
"type": "string"
},
"description": "mRNA 列表"
},
"protein": {
"type": "array",
"items": {
"type": "string"
},
"description": "蛋白列表"
},
"metabolite": {
"type": "array",
"items": {
"type": "string"
},
"description": "代谢物列表"
},
"pathway_map": {
"type": "string",
"description": "通路映射方式"
}
}
}【置换检验】 数据不满足 t 检验前提、样本很小、要做非参数精确置换检验 返回:置换 P 值 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| group1* | array | 第一组数值 |
| group2* | array | 第二组数值 |
| n_perm | integer | 置换次数,默认 10000 |
{
"type": "object",
"required": [
"group1",
"group2"
],
"properties": {
"group1": {
"type": "array",
"items": {
"type": "string"
},
"description": "第一组数值"
},
"group2": {
"type": "array",
"items": {
"type": "string"
},
"description": "第二组数值"
},
"n_perm": {
"type": "integer",
"description": "置换次数,默认 10000"
}
}
}【PICO 精炼】 用户给了研究问题,要拆成规范的 PICO 返回:结构化 PICO (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| pico* | object | 结构化 PICO,如 {"P":"ICU 脓毒症患者","I":"乳酸清除率","C":"常规","O":"28天病死率"} |
{
"type": "object",
"required": [
"pico"
],
"properties": {
"pico": {
"type": "object",
"description": "结构化 PICO,如 {\"P\":\"ICU 脓毒症患者\",\"I\":\"乳酸清除率\",\"C\":\"常规\",\"O\":\"28天病死率\"}"
}
}
}【PK/PD 评估】 用户要做药代药效评估、剂量推算 返回:PK/PD 评估结果 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| drug_name* | string | 药物名称 |
| pk_params | object | 药代参数 |
{
"type": "object",
"required": [
"drug_name"
],
"properties": {
"drug_name": {
"type": "string",
"description": "药物名称"
},
"pk_params": {
"type": "object",
"description": "药代参数"
}
}
}【PK/PD 模拟】 用户要模拟给药方案的浓度-效应过程 返回:模拟曲线与参数 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| drug_name* | string | 药物名 |
| pd_params | object | 药效参数 |
| pk_params | object | 药代参数 |
{
"type": "object",
"required": [
"drug_name"
],
"properties": {
"drug_name": {
"type": "string",
"description": "药物名"
},
"pd_params": {
"type": "object",
"description": "药效参数"
},
"pk_params": {
"type": "object",
"description": "药代参数"
}
}
}【蛋白互作网络】 用户有一批蛋白要做 PPI 网络、找 hub 返回:互作网络与 hub 蛋白 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| species | string | 物种:human/mouse |
| protein_list* | array | 蛋白/基因列表 |
| confidence_threshold | number | 置信度阈值 |
{
"type": "object",
"required": [
"protein_list"
],
"properties": {
"species": {
"type": "string",
"description": "物种:human/mouse"
},
"protein_list": {
"type": "array",
"items": {
"type": "string"
},
"description": "蛋白/基因列表"
},
"confidence_threshold": {
"type": "number",
"description": "置信度阈值"
}
}
}【PCR 引物设计】 用户给了 DNA 序列或基因,要设计 PCR 引物 返回:候选引物对与 Tm/GC (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| gene | string | 基因名(无序列时用) |
| sequence | string | 模板 DNA 序列 |
{
"type": "object",
"required": [],
"properties": {
"gene": {
"type": "string",
"description": "基因名(无序列时用)"
},
"sequence": {
"type": "string",
"description": "模板 DNA 序列"
}
}
}【引物 Tm 计算】 用户给了引物序列要算 Tm、GC、二聚体 返回:Tm、GC、二聚体 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 引物序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "引物序列"
}
}
}【PRISMA 流程图】 用户做系统评价/Meta 分析,要画文献筛选流程图 返回:PRISMA 流程图 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| title | string | 研究题目 |
| records_screened | integer | 筛读记录数 |
| studies_included | integer | 最终纳入研究数 |
| records_identified | integer | 检索记录数 |
{
"type": "object",
"required": [],
"properties": {
"title": {
"type": "string",
"description": "研究题目"
},
"records_screened": {
"type": "integer",
"description": "筛读记录数"
},
"studies_included": {
"type": "integer",
"description": "最终纳入研究数"
},
"records_identified": {
"type": "integer",
"description": "检索记录数"
}
}
}【蛋白酶解模拟】 用户要选酶做质谱前处理、算肽段 返回:肽段清单 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| enzymes | array | 酶:Trypsin/Lys-C |
| sequence* | string | 蛋白序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"enzymes": {
"type": "array",
"items": {
"type": "string"
},
"description": "酶:Trypsin/Lys-C"
},
"sequence": {
"type": "string",
"description": "蛋白序列"
}
}
}【蛋白富集分析】 用户有一批蛋白要做 GO/KEGG 富集 返回:富集通路 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| species | string | 物种 |
| protein_list* | array | 蛋白列表 |
| enrichment_type | string | GO/KEGG/Reactome |
{
"type": "object",
"required": [
"protein_list"
],
"properties": {
"species": {
"type": "string",
"description": "物种"
},
"protein_list": {
"type": "array",
"items": {
"type": "string"
},
"description": "蛋白列表"
},
"enrichment_type": {
"type": "string",
"description": "GO/KEGG/Reactome"
}
}
}【蛋白理化性质】 用户给蛋白序列要算分子量/等电点/疏水性 返回:MW/pI/GRAVY 等 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 蛋白序列 |
| protein_name | string | 蛋白名 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "蛋白序列"
},
"protein_name": {
"type": "string",
"description": "蛋白名"
}
}
}【翻译后修饰分析】 用户要预测磷酸化/糖基化等修饰位点 返回:修饰位点预测 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 蛋白序列 |
| modifications | array | 修饰类型:phosphorylation/glycosylation |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "蛋白序列"
},
"modifications": {
"type": "array",
"items": {
"type": "string"
},
"description": "修饰类型:phosphorylation/glycosylation"
}
}
}【倾向性评分匹配 PSM】 用户要做 PSM、倾向评分匹配、1:1 匹配基线 返回:匹配后样本与均衡性 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data | array | 样本数据 |
| ratio | number | 匹配比例,默认 1 |
| treatment | string | 处理变量列名 |
| covariates | array | 协变量列名 |
{
"type": "object",
"required": [],
"properties": {
"data": {
"type": "array",
"items": {
"type": "string"
},
"description": "样本数据"
},
"ratio": {
"type": "number",
"description": "匹配比例,默认 1"
},
"treatment": {
"type": "string",
"description": "处理变量列名"
},
"covariates": {
"type": "array",
"items": {
"type": "string"
},
"description": "协变量列名"
}
}
}【PubMed 检索】 用户要查 PubMed、要最新文献 返回:PubMed 文献列表 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| q* | string | 检索式(英文) |
| year_start | integer | 起始年份 |
| max_results | integer | 条数 |
{
"type": "object",
"required": [
"q"
],
"properties": {
"q": {
"type": "string",
"description": "检索式(英文)"
},
"year_start": {
"type": "integer",
"description": "起始年份"
},
"max_results": {
"type": "integer",
"description": "条数"
}
}
}【qPCR 相对表达量(2^-ΔΔCt)】 用户给出目的基因与内参基因的 Ct 值(对照组/处理组各若干重复),要算 ΔCt、ΔΔCt、2^-ΔΔCt 相对表达量或倍数、并做组间检验时,必须走这个动作。🔴 严禁自己心算 ΔΔCt 与倍数:校准必须用**对照组 ΔCt 的均值**(不是某一个对照样本的值,用了单样本会整体偏高/偏低);且统计检验要由引擎判断(组内零方差时 Welch t 会给出虚高的 P 值,引擎会改用 Mann-Whitney)。返回里 stat_method / tested_variable 必须如实转述。Ct 值用逗号分隔,两组与各自内参一一对应。 返回:fc(2^-ΔΔCt 相对倍数,已按对照组 ΔCt 均值校准)、fold_values、p_value、stat_method(welch_ttest_on_dct / student_ttest_on_dct / mann_whitney)、tested_variable(固定 delta_ct)、dct、significance、outlier_check (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| gene | string | 目的基因名,如 NLRP3 / TREM2 |
| exp_ct* | string | 处理组目的基因 Ct,逗号分隔,如 20.1,20.4,20.2 |
| ctrl_ct* | string | 对照组目的基因 Ct,逗号分隔,如 22.5,22.8,22.6 |
| ref_gene | string | 内参基因名,默认 GAPDH |
| exp_ref_ct* | string | 处理组内参基因 Ct,逗号分隔,与 exp_ct 一一对应,如 18.2,18.5,18.3 |
| ctrl_ref_ct* | string | 对照组内参基因 Ct,逗号分隔,与 ctrl_ct 一一对应,如 18.3,18.1,18.4 |
{
"type": "object",
"required": [
"ctrl_ct",
"ctrl_ref_ct",
"exp_ct",
"exp_ref_ct"
],
"properties": {
"gene": {
"type": "string",
"description": "目的基因名,如 NLRP3 / TREM2"
},
"exp_ct": {
"type": "string",
"description": "处理组目的基因 Ct,逗号分隔,如 20.1,20.4,20.2"
},
"ctrl_ct": {
"type": "string",
"description": "对照组目的基因 Ct,逗号分隔,如 22.5,22.8,22.6"
},
"ref_gene": {
"type": "string",
"description": "内参基因名,默认 GAPDH"
},
"exp_ref_ct": {
"type": "string",
"description": "处理组内参基因 Ct,逗号分隔,与 exp_ct 一一对应,如 18.2,18.5,18.3"
},
"ctrl_ref_ct": {
"type": "string",
"description": "对照组内参基因 Ct,逗号分隔,与 ctrl_ct 一一对应,如 18.3,18.1,18.4"
}
}
}【随机化分组】 用户要随机分组、随机数字表、区组随机、分层随机 返回:分组结果表 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| seed | integer | 随机种子(固定才可复现) |
| groups* | array | 分组名称数组,如 ["干预组","对照组"](必须是数组,不是数字) |
| method | string | simple 简单 / block 区组 / stratified 分层 / minimization 最小化 |
| block_size | integer | 区组长度(block 方法用) |
| n_subjects* | integer | 总样本量,如 60 |
{
"type": "object",
"required": [
"n_subjects",
"groups"
],
"properties": {
"seed": {
"type": "integer",
"description": "随机种子(固定才可复现)"
},
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组名称数组,如 [\"干预组\",\"对照组\"](必须是数组,不是数字)"
},
"method": {
"type": "string",
"description": "simple 简单 / block 区组 / stratified 分层 / minimization 最小化"
},
"block_size": {
"type": "integer",
"description": "区组长度(block 方法用)"
},
"n_subjects": {
"type": "integer",
"description": "总样本量,如 60"
}
}
}【相对危险度 RR】 用户给的是四格表人数,要 RR 或 OR 及 CI 返回:RR、OR、95%CI (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| exposed_total* | integer | 暴露组总数 |
| exposed_events* | integer | 暴露组事件数 |
| unexposed_total* | integer | 非暴露组总数 |
| unexposed_events* | integer | 非暴露组事件数 |
{
"type": "object",
"required": [
"exposed_events",
"exposed_total",
"unexposed_events",
"unexposed_total"
],
"properties": {
"exposed_total": {
"type": "integer",
"description": "暴露组总数"
},
"exposed_events": {
"type": "integer",
"description": "暴露组事件数"
},
"unexposed_total": {
"type": "integer",
"description": "非暴露组总数"
},
"unexposed_events": {
"type": "integer",
"description": "非暴露组事件数"
}
}
}【RNA-seq 差异分析】 用户有 RNA-seq 表达矩阵要做差异基因 返回:差异基因清单 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| groups | array | 分组 |
| log2fc | number | log2FC 阈值 |
| data_json* | object | 表达矩阵 |
| fdr_threshold | number | FDR |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"groups": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组"
},
"log2fc": {
"type": "number",
"description": "log2FC 阈值"
},
"data_json": {
"type": "object",
"description": "表达矩阵"
},
"fdr_threshold": {
"type": "number",
"description": "FDR"
}
}
}【ROC 曲线与 AUC】 用户有金标准标签和预测分数,要算 AUC、画 ROC 返回:AUC、最优阈值、敏感度特异度 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| labels* | array | 真实标签 0/1 |
| scores* | array | 预测分数/概率 |
{
"type": "object",
"required": [
"labels",
"scores"
],
"properties": {
"labels": {
"type": "array",
"items": {
"type": "string"
},
"description": "真实标签 0/1"
},
"scores": {
"type": "array",
"items": {
"type": "string"
},
"description": "预测分数/概率"
}
}
}【生成可复现 SAS/R 代码】 用户要把分析结果导出成 R / SAS / Python 代码复现 返回:推荐统计方法与代码 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| design* | string | 研究设计 |
| n_groups | integer | 组数 |
| outcome_type | string | 结局类型 |
{
"type": "object",
"required": [
"design"
],
"properties": {
"design": {
"type": "string",
"description": "研究设计"
},
"n_groups": {
"type": "integer",
"description": "组数"
},
"outcome_type": {
"type": "string",
"description": "结局类型"
}
}
}【单细胞分析】 用户有单细胞数据要做聚类降维 返回:聚类结果与 UMAP (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| n_pcs | integer | 主成分数 |
| data_json* | object | 单细胞矩阵 |
| min_genes | integer | 最少基因数 |
| max_mito_pct | number | 线粒体比例上限 |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"n_pcs": {
"type": "integer",
"description": "主成分数"
},
"data_json": {
"type": "object",
"description": "单细胞矩阵"
},
"min_genes": {
"type": "integer",
"description": "最少基因数"
},
"max_mito_pct": {
"type": "number",
"description": "线粒体比例上限"
}
}
}【单细胞细胞类型注释】 用户要自动注释细胞类型 返回:细胞类型注释 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【单细胞 Marker 基因】 用户要找各细胞群的 marker 基因 返回:各群 marker (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data_json* | object | 必填。数据对象(不是字符串)。矩阵形式:{"matrix": [[行1各列数值...], [行2...]], "rows": ["基因/特征名"], "cols": ["样本名"]};组学结果另可带 {"groups": ["组A","组A","组B"...]}(与 cols 一一对应)、{"p_values": [...], "log2fc": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json |
{
"type": "object",
"required": [
"data_json"
],
"properties": {
"data_json": {
"type": "object",
"description": "必填。数据对象(不是字符串)。矩阵形式:{\"matrix\": [[行1各列数值...], [行2...]], \"rows\": [\"基因/特征名\"], \"cols\": [\"样本名\"]};组学结果另可带 {\"groups\": [\"组A\",\"组A\",\"组B\"...]}(与 cols 一一对应)、{\"p_values\": [...], \"log2fc\": [...]}(若已有差异结果)。实测缺此字段会被拒:Field required: data_json"
}
}
}【酶切位点分析】 用户问序列上有哪些酶切位点、能不能切开 返回:酶切位点清单 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 核酸序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "核酸序列"
}
}
}【ORF 预测】 用户要在序列里找开放阅读框 返回:ORF 列表与位置 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 核酸序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "核酸序列"
}
}
}【质粒图谱】 用户要画质粒图谱、看载体结构 返回:质粒图谱 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 质粒序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "质粒序列"
}
}
}【序列翻译】 用户贴了核酸序列要翻译成氨基酸 返回:氨基酸序列 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| sequence* | string | 核酸序列 |
{
"type": "object",
"required": [
"sequence"
],
"properties": {
"sequence": {
"type": "string",
"description": "核酸序列"
}
}
}【SMD 均衡性检查】 用户做 PSM/IPTW 后要检查协变量均衡(标准化均数差) 返回:各协变量 SMD 与均衡判断 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| treatment* | array | 分组 0/1 |
| covariates* | array | 协变量矩阵:每行一个样本、每列一个变量 |
{
"type": "object",
"required": [
"covariates",
"treatment"
],
"properties": {
"treatment": {
"type": "array",
"items": {
"type": "string"
},
"description": "分组 0/1"
},
"covariates": {
"type": "array",
"items": {
"type": "string"
},
"description": "协变量矩阵:每行一个样本、每列一个变量"
}
}
}【统计中心(97 个统计工具统一入口)】 用户要跑的统计检验没被其他专门工具覆盖时用这个:比较差异、找关系、看生存、做诊断、队列预处理、综合证据六类共 97 个工具(Kappa、Bland-Altman、ICC、VIF、PCA、聚类、Friedman、McNemar、游程、符号检验、趋势卡方、Meta 各件等都从这里走) 返回:该统计工具的完整结果(含统计量、P 值、95%CI、解读) (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| payload | object | 该工具需要的数据字段(不要把数据包成字符串)。常见字段:group1/group2(两组数值数组)、groups({组名:[数值]})、values、x/y、time/event、gold/score、es/se(效应量与标准误)、m1/m2、r1/r2、row_var/col_var、features/label |
| tool_id* | string | 统计工具 id,必须是下列之一:比较差异[ttest,welch,mwu,wilcoxon,anova,welchAnova,kw,friedman,rmAnova,ancova,chi2,fisher,mcnemar,cochranQ,posthoc,multicorr,homogeneity,normality,brunner,signTest,runsTest,effectSize,pInverse]、找关系[pearson,spearman,kendall,linearReg,multiLinear,logistic,logisticModel,vif,regDiag,icc,blan |
{
"type": "object",
"required": [
"tool_id"
],
"properties": {
"payload": {
"type": "object",
"description": "该工具需要的数据字段(不要把数据包成字符串)。常见字段:group1/group2(两组数值数组)、groups({组名:[数值]})、values、x/y、time/event、gold/score、es/se(效应量与标准误)、m1/m2、r1/r2、row_var/col_var、features/label"
},
"tool_id": {
"type": "string",
"description": "统计工具 id,必须是下列之一:比较差异[ttest,welch,mwu,wilcoxon,anova,welchAnova,kw,friedman,rmAnova,ancova,chi2,fisher,mcnemar,cochranQ,posthoc,multicorr,homogeneity,normality,brunner,signTest,runsTest,effectSize,pInverse]、找关系[pearson,spearman,kendall,linearReg,multiLinear,logistic,logisticModel,vif,regDiag,icc,blan"
}
}
}【生存分析(KM/Cox 全流程)】 用户有随访时间与结局事件,要做生存分析、算 HR、出生存曲线 返回:生存曲线、中位生存、HR、log-rank P (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| data* | array | 逐行数据,每行含时间/事件/分组/协变量 |
| time_col | string | 随访时间列名 |
| event_col | string | 结局事件列名(1=事件) |
| group_col | string | 分组列名 |
| analysis_type | string | km 或 cox |
| covariate_cols | array | Cox 协变量列名 |
{
"type": "object",
"required": [
"data"
],
"properties": {
"data": {
"type": "array",
"items": {
"type": "string"
},
"description": "逐行数据,每行含时间/事件/分组/协变量"
},
"time_col": {
"type": "string",
"description": "随访时间列名"
},
"event_col": {
"type": "string",
"description": "结局事件列名(1=事件)"
},
"group_col": {
"type": "string",
"description": "分组列名"
},
"analysis_type": {
"type": "string",
"description": "km 或 cox"
},
"covariate_cols": {
"type": "array",
"items": {
"type": "string"
},
"description": "Cox 协变量列名"
}
}
}【生成 Table 1(基线特征表)】 用户有数据文件,要出论文第一张基线表 Table 1 返回:Table 1 表格 + 可选段落 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| file_id | integer | 数据文件 id(可不填,助手会自动用你最近上传的数据) |
| test_method | string | auto/t/nonparametric/chi2 |
| group_variable* | string | 分组变量列名 |
| generate_paragraph | boolean | 是否同时生成论文表述段落 |
| variables_to_include | array | 纳入的变量列名 |
{
"type": "object",
"required": [
"group_variable"
],
"properties": {
"file_id": {
"type": "integer",
"description": "数据文件 id(可不填,助手会自动用你最近上传的数据)"
},
"test_method": {
"type": "string",
"description": "auto/t/nonparametric/chi2"
},
"group_variable": {
"type": "string",
"description": "分组变量列名"
},
"generate_paragraph": {
"type": "boolean",
"description": "是否同时生成论文表述段落"
},
"variables_to_include": {
"type": "array",
"items": {
"type": "string"
},
"description": "纳入的变量列名"
}
}
}【分组 Table 1(亚组表)】 用户要按某个变量做分层的基线表 返回:亚组表 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| file_id | integer | 数据文件 id |
| group_variable* | string | 分组变量 |
| subgroup_variable* | string | 分层变量 |
| variables_to_include | array | 纳入变量 |
{
"type": "object",
"required": [
"group_variable",
"subgroup_variable"
],
"properties": {
"file_id": {
"type": "integer",
"description": "数据文件 id"
},
"group_variable": {
"type": "string",
"description": "分组变量"
},
"subgroup_variable": {
"type": "string",
"description": "分层变量"
},
"variables_to_include": {
"type": "array",
"items": {
"type": "string"
},
"description": "纳入变量"
}
}
}【肿瘤微环境分析】 用户有 TILs/CD8/PD-L1 数据要评估肿瘤微环境 返回:微环境分型与预后提示 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| tils_pct | number | TILs 占比 |
| cd8_count | number | CD8+ 细胞数 |
| pd_l1_pos | number | PD-L1 阳性率 |
| cancer_type | string | 癌种 |
{
"type": "object",
"required": [],
"properties": {
"tils_pct": {
"type": "number",
"description": "TILs 占比"
},
"cd8_count": {
"type": "number",
"description": "CD8+ 细胞数"
},
"pd_l1_pos": {
"type": "number",
"description": "PD-L1 阳性率"
},
"cancer_type": {
"type": "string",
"description": "癌种"
}
}
}【毒理评估】 用户要评估候选物的毒性风险(hERG/DILI/遗传毒) 返回:毒性风险分级 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| logP | number | logP |
| hERG_IC50_nM | number | hERG IC50 (nM) |
| daily_dose_mg | number | 日剂量 (mg) |
| expected_Cmax_nM | number | 预期 Cmax (nM) |
{
"type": "object",
"required": [],
"properties": {
"logP": {
"type": "number",
"description": "logP"
},
"hERG_IC50_nM": {
"type": "number",
"description": "hERG IC50 (nM)"
},
"daily_dose_mg": {
"type": "number",
"description": "日剂量 (mg)"
},
"expected_Cmax_nM": {
"type": "number",
"description": "预期 Cmax (nM)"
}
}
}【毒性预测(hERG/DILI)】 用户要评估化合物心脏/肝毒性风险 返回:毒性风险分级 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| smiles* | string | 化合物 SMILES |
| endpoints | array | 终点:hERG/DILI/Ames |
{
"type": "object",
"required": [
"smiles"
],
"properties": {
"smiles": {
"type": "string",
"description": "化合物 SMILES"
},
"endpoints": {
"type": "array",
"items": {
"type": "string"
},
"description": "终点:hERG/DILI/Ames"
}
}
}【多组两两比较(Tukey HSD)】 用户做了 ANOVA 后要多组两两比较 返回:两两比较 P 值 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| groups* | object | {组名: [数值...]} |
{
"type": "object",
"required": [
"groups"
],
"properties": {
"groups": {
"type": "object",
"description": "{组名: [数值...]}"
}
}
}【双因素方差分析】 用户有两个因素要看主效应与交互作用 返回:主效应与交互 (真实算法引擎执行,非大模型生成)
| Argument | Type | Description |
|---|---|---|
| values | array | 观测值 |
| group_a* | array | 因素 A 水平 |
| group_b* | array | 因素 B 水平 |
{
"type": "object",
"required": [
"group_a",
"group_b"
],
"properties": {
"values": {
"type": "array",
"items": {
"type": "string"
},
"description": "观测值"
},
"group_a": {
"type": "array",
"items": {
"type": "string"
},
"description": "因素 A 水平"
},
"group_b": {
"type": "array",
"items": {
"type": "string"
},
"description": "因素 B 水平"
}
}
}列出三域医研全部真算法动作的摘要(id / 名称 / 适用场景 / 产出 / 必填参数 / 计费)。仅在不确定该用哪个动作时调用;已知动作 id 可直接调用同名工具。
No arguments.
{
"type": "object",
"required": [],
"properties": {}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 6, 2026.
{
"type": "object",
"required": [
"data",
"formula",
"subject"
],
"properties": {
"data": {
"type": "array",
"items": {
"type": "string"
},
"description": "长格式数据:每行一次观测 [{受试者ID, 分组, 时点, 结局}, ...];也接受列式 {列名: [值, ...]}。注意必须是**长表**(每行一次测量),不是每人一行多列"
},
"reml": {
"type": "boolean",
"description": "默认 true(REML)"
},
"formula": {
"type": "string",
"description": "R 风格公式,如 '收缩压 ~ 分组 * 时点'(组×时间交互);左端=结局"
},
"subject": {
"type": "string",
"description": "受试者/病案号列名(随机截距的分组依据)"
},
"reference": {
"type": "object",
"description": "参照水平,如 {\"分组\": \"对照组\"} —— 报告表里「与对照组相比」必须是它"
},
"re_formula": {
"type": "string",
"description": "随机效应,默认 '1'(随机截距);随机斜率用 '1 + 时点'"
},
"categorical": {
"type": "array",
"items": {
"type": "string"
},
"description": "显式声明哪些列按分类处理(如 [\"分组\"]);数值编码的 0/1 分组必须声明,否则会被当连续变量"
},
"time_values": {
"type": "object",
"description": "把字符串时点映射成数值以估线性趋势,如 {\"时点\": {\"基线\":0,\"4周\":1,\"12周\":2}}"
}
}
}