虚拟细胞建模里程碑：粒曼生物近期将正式发布 “ELEM‑1200RS”，面向虚拟细胞挑战赛的大规模 bulk 扰动转录组图谱，为基因扰动建模、虚拟细胞研究及药物靶点研发带来重大跨越。

 

A milestone for virtual‑cell builders: We have released ELEM‑1200RS, a large‑scale bulk RNA-seq perturbation atlas built for the Virtual Cell Challenge. This is a major leap forward for genetic‑perturbation modelling, virtual cell research, and therapeutic target discovery.

 

背景材料：AIVC专题｜AIVC进入数据竞赛：谁能提供可验证的细胞扰动数据？

 

ELEM-1200RS数据集下载说明

一、数据集简介

ELEM Biotech（粒曼生物）虚拟细胞扰动转录组学数据集 "ELEM-1200RS" 包含约1,200个 bulk RNA-seq样本，来源于63种野生型细胞系背景，敲低实验共靶向914个独特基因。基因表达数据以 TPM 标准化表达矩阵形式提供，采用 .h5ad 格式，与虚拟细胞挑战赛（Virtual Cell Challenge）所使用的数据结构保持一致。

 

数据集通过Hugging Face平台发布，优先访问期间采用“访问门控方式”管理下载权限。

数据集访问地址：https://huggingface.co/datasets/ELEM-Bio/ELEM-1200RS

注：公开发布的数据集不提供样本编号与扰动靶基因、细胞系名称的映射关系。如需获取完整注释元数据，请向ELEM邮箱sales@elem-bio.com发送申请，经确认后我们将提供相应注释信息。

图1 阵列文库Arrayed Library基因扰动数据集与混合文库Pooled Library基因扰动数据集的区别

 

二、优先访问期间的下载流程

（2026年9月30日至2026年10月12日，北京时间 UTC+8）

优先访问期间，数据集向已签订《AIVC数据优先访问与使用协议》的指定合作团队免费提供优先访问和使用权，下载流程如下：

1. 登录Hugging Face平台，访问上述链接，或检索数据集名称"ELEM-1200RS"进入数据集主页；

2. 在数据集的"Dataset card"页面，请点击"Agree and send request to access repo"，确认提交数据访问申请；

3. 请将您的 Hugging Face 用户名及注册邮箱等信息发送至我方邮箱sales@elem-bio.com，邮件模版见 【四、申请邮件模板（1.优先访问期间）】；

4. 收到邮件后，将由 ELEM 完成您的数据访问权限释放，同时提供与表达矩阵配套的完整注释元数据。您将收到Hugging Face系统发送的通知邮件，届时即可登录平台完成数据集下载。

备注：申请优先使用数据的团队，请发邮件至yongbo_cheng@elem-bio.cn，获取《AIVC数据优先访问与使用协议》文件，并在2026年9月24日至2026年9月30日期间内完成协议签订。

 

三、全面公开发布的下载流程

（自 2026年10月13日起，北京时间 UTC+8）

自2026年10月13日起，ELEM将在全球范围内公开发布该数据集，用户无需再申请访问权限，下载流程如下：

1. 登录Hugging Face平台，访问上述链接，或检索数据集名称 "ELEM-1200RS" 进入数据集主页；

2. 在数据集的 "Files and versions" 页面即可直接下载对应数据集；

3. 请将您的Hugging Face用户名及注册邮箱等信息发送至我方邮箱sales@elem-bio.com，邮件模版见【四、申请邮件模板（2.公开发布期间）】；

4. 收到邮件后，ELEM将提供与表达矩阵配套的完整注释元数据。 

 

四、申请邮件模板

1. 优先访问期间

收件人：sales@elem-bio.com

邮件主题：ELEM-1200RS注释信息申请

邮件正文：

ELEM Biotech 团队，

我方已签署《AIVC数据优先访问与使用协议》，并已在 Hugging Face 平台提交虚拟细胞扰动转录组学数据集ELEM‑1200RS 下载申请。现申请获取扰动靶基因名称及细胞系名称映射注释信息，拟用于（请简要填写研究/使用目的）。

申请人姓名：

单位/团队：

Hugging Face 用户名：

联系邮箱：

联系电话：

 

2. 公开发布期间

收件人：sales@elem-bio.com

邮件主题：ELEM-1200RS注释信息申请

邮件正文：

ELEM Biotech 团队，

我方已于Hugging Face平台下载虚拟细胞扰动转录组学数据集ELEM‑1200RS。现申请获取扰动靶基因名称及细胞系名称映射注释信息，拟用于（请简要填写研究/使用目的）。

申请人姓名：

单位/团队：

Hugging Face 用户名：

联系邮箱：

联系电话：

 

 

ELEM-1200RS Dataset Download Instructions

 

1. Dataset Overview

The ELEM Biotech Virtual Cell Perturbation Transcriptomics Dataset ("ELEM-1200RS") comprises approximately 1,200 bulk RNA-seq samples derived from 63 wild-type cell lines, with knockdown experiments targeting 914 unique genes. Gene expression data are provided as TPM-normalized expression matrices in .h5ad format, consistent with the data structure adopted by the Virtual Cell Challenge.

The dataset is released on the Hugging Face platform. During the priority access period, download permissions are administered through gated access.

Dataset URL: https://huggingface.co/datasets/ELEM-Bio/ELEM-1200RS

 

Note: The publicly released dataset does not include the mapping between sample IDs and the corresponding target-gene names or cell line names. To obtain the complete annotation metadata, please submit your request to ELEM via email: sales@elem-bio.com.The corresponding annotation information will be provided upon confirmation.

 

2. Download Procedure During the Priority Window 

(September 30, 2026 to October 12, 2026, Beijing Time, UTC+8)

During the priority window, the dataset is made available free of charge for priority access and use to designated collaboratingteams that have executed the Dataset Priority Use Agreement. The download procedure is as follows:

1. Log in to the Hugging Face platform. Navigate to the link provided above, or search for the dataset name 'ELEM‑1200RS' to access the dataset homepage;

2. On the 'Dataset card'page, click 'Agree and send request to access repo'to submit your data access application;

3. Send your Hugging Face username, registered email address, and other relevant information to sales@elem-bio.com. An email template is provided in Section 4.1 (Priority Window);

4. Upon receipt of your email, ELEM will grant you dataset-access permissions and provide the annotation metadata accompanying the expression matrix. Hugging Face will notify you when access is granted, after which you may log in to download the dataset.

Note:Teams wishing to apply for priority data access and usage should email yongbo_cheng@elem-bio.cn to obtain the "AIVC Data Priority Access and Usage Agreement" and complete the signing of the agreement between September 24, 2026, and September 30, 2026.

 

3. Download Procedure Following the Global Public Release

 (Effective October 13, 2026，Beijing Time, UTC+8)

Effective October 13, 2026, ELEM will publicly release the dataset worldwide. The download procedure is as follows:

1. Log in to the Hugging Face platform. Navigate to the link provided above, or search for the dataset name 'ELEM‑1200RS' to access the dataset homepage;

2. You may directly download the dataset files from the 'Files and versions'page;

3. Send your Hugging Face username, registered email address, and other relevant information to sales@elem-bio.com. An email template is provided in Section 4.2 (Public Release Period);

4. Upon receipt of your email, ELEM will provide the annotation metadata accompanying the expression matrix.

 

4. Email Templates

 

1. Priority Window

Recipient: sales@elem-bio.com

Subject: Request for ELEM-1200RS metadata

Email Body:

Dear ELEM Biotech Team,

We have executed the AIVC DATA PRIORITY ACCESS AND USE AGREEMENT and have submitted a download request for theVirtual Cell Perturbation Transcriptomics Dataset 'ELEM-1200RS' on the Hugging Face.

We hereby request the corresponding mapping metadata for the perturbed target-gene names and cell line names. This information will be used for the following research purpose: (please briefly state your research/use purpose).

 

Applicant name:

Institution/Team:

Hugging Face username:

Contact email:

telephone：

Best regards

 

2. Public Release Period

Recipient: sales@elem-bio.com

Subject: Request for ELEM-1200RS metadata

Email Body:

Dear ELEM Biotech Team,

We have downloaded the Virtual Cell Perturbation Transcriptomics Dataset 'ELEM-1200RS' from Hugging Face. We hereby request the corresponding mapping metadata for the perturbed target-gene names and cell line names. This information will be used for the following research purpose: (please briefly state your research/use purpose).

 

Applicant name:

Institution/Team:

Hugging Face username:

Contact email:

telephone：

Best regards