NEW YORK – Illumina and Tempus AI are teaming up to improve artificial intelligence-based tools that they hope will be used for interpretation of molecular tests.

纽约——Illumina和Tempus AI正在合作改进基于人工智能的工具，他们希望这些工具能用于分子测试的解读。

Under a three-year deal

三年协议之下

announced

宣布

earlier this week

本周早些时候

, Illumina has licensed access to Tempus' multimodal data platform, which includes not only DNA and RNA data but also clinical data, pathology slides, and radiology scans. The data are de-identified, Tempus said, and the deal is not exclusive.

，Illumina已获得授权访问Tempus的多模态数据平台，该平台不仅包括DNA和RNA数据，还包括临床数据、病理切片和放射学扫描。Tempus表示，这些数据是去标识化的，并且该协议并非独家。

Illumina plans to use that access to help improve its AI algorithms, including

Illumina计划利用这一访问权限来帮助改进其AI算法，包括

PrimateAI-3D

PrimateAI-3D

, a semi-supervised 3D-convoluational neural network to predict variant pathogenicity, and SpliceAI, for interpreting noncoding mutations. Insights from the collaboration will be shared by both firms, an Illumina spokesperson said. Financial and other details were not disclosed.

，一种半监督的3D卷积神经网络，用于预测变异致病性，以及SpliceAI，用于解读非编码突变。Illumina发言人表示，合作的见解将由两家公司共享。财务和其他细节未披露。

The firms said they hope to generate new evidence supporting the clinical value of sequencing, which 'will be used to build evidence packages needed to standardize use of comprehensive genomic profiling and other molecular testing across all major diseases.'

这些公司表示，他们希望生成新的证据，以支持测序的临床价值，“这些证据将用于构建必要的证据包，以标准化在所有主要疾病中使用全面的基因组分析和其他分子测试。”

Getting PrimateAI-3D to work better might help drive its adoption. 'There are a ton of predictor algorithms, but this is not one of the widely used ones,' Heidi Rehm, chief genomics officer at the Center for Genomic Medicine at Massachusetts General Hospital and an expert on variant interpretation, said in an email.

让PrimateAI-3D更好地发挥作用可能会促进其应用。马萨诸塞州总医院基因组医学中心的首席基因组官、变异解读专家海蒂·雷姆在一封电子邮件中表示：“目前存在许多预测算法，但这并不是广泛使用的其中之一。”

.

。

Illumina's January

因美纳的一月

deal

交易

with Nvidia provides the proper computing resources for developing better AI tools, said Kyle Farh, an Illumina VP in charge of AI, while 'Tempus brings in the datasets needed for training.' These are the key pieces the firm sees powering its push into this space.

英伟达提供了开发更好的人工智能工具所需的适当计算资源，Illumina 负责人工智能的副总裁凯尔·法尔表示，而“Tempus 提供了训练所需的数据集。” 这些是该公司认为推动其进入这一领域的关键要素。

'Tempus has a longstanding relationship with Illumina as a user of their technology,' Tempus Chief Commercial Officer Terron Bruner said in an email. 'We're looking forward to expanding our relationship through this new collaboration.'

“Tempus与Illumina作为其技术用户有着长期的合作关系，”Tempus首席商务官Terron Bruner在一封电子邮件中表示，“我们期待通过这一新的合作来扩展我们的关系。”

As Illumina's clinical customers move to whole-genome sequencing, they're finding an 'increased interpretation burden,' Farh said. A patient genome might yield 4 million variants, with the vast majority being variants of unknown significance (VUS). 'We want to understand more about how these [VUS] impact health,' he said.

法尔表示，随着Illumina的临床客户转向全基因组测序，他们发现“解读负担增加了”。一个患者的基因组可能会产生400万个变异，其中绝大多数是意义不明的变异（VUS）。他说：“我们希望更多地了解这些[意义不明的变异]如何影响健康。”

.

。

Some of Illumina's algorithms, including PrimateAI-3D, attempt to predict variant function from sequence. PrimateAI-3D looks at genomic data from over

Illumina的一些算法，包括PrimateAI-3D，试图从序列预测变异功能。PrimateAI-3D分析了来自超过

200 nonhuman primate species

200种非人灵长类动物物种

to see where mutations are tolerated and where they aren't, as well as 3D protein modeling to predict the function of a human VUS.

观察哪些位置的突变可以被容忍，哪些不能，以及使用3D蛋白质建模来预测人类VUS的功能。

If the predictions are improved, sequences with a VUS could be shown to be similar in function to an established target for therapy, which would potentially increase the number of patients who could benefit from that drug.

如果预测得到改进，带有 VUS 的序列可能在功能上被证明与已确立的治疗靶点相似，这将有可能增加受益于该药物的患者数量。

PrimateAI-3D and SpliceAI are available through public code repositories, Farh noted, but they're also integrated into Illumina software products such as Emedgene variant interpretation and Illumina Connected Insights.

Farh 指出，PrimateAI-3D 和 SpliceAI 可通过公共代码库获取，但它们也被整合到 Illumina 的软件产品中，例如 Emedgene 变异解读和 Illumina Connected Insights。

Illumina and Tempus will initially focus on cancer but are hoping to expand the collaboration to many other areas of medicine.

Illumina和Tempus将首先专注于癌症领域，但希望将合作扩展到医学的许多其他领域。

'The algorithms we develop are for general genome interpretation,' Farh said, noting that they work for both rare diseases and cancer. 'This specific partnership starts with oncology data first, because that's what [Tempus has].'

“我们开发的算法适用于一般的基因组解读，”法尔表示，他指出这些算法既适用于罕见病，也适用于癌症。“这个特定的合作关系首先从肿瘤学数据开始，因为这是[Tempus公司]所拥有的。”

'In the era of true precision medicine, every patient who is battling complex disease should be routed to the optimal therapy based on molecular insights,' Everett Cunningham, chief commercial officer of Illumina, said in a statement. 'We envision a world where the full range of molecular profiling is available as part of the standard of care — not just in cancer, but in cardiology, neurology, immunology, and every other category of disease.'.

“在真正精准医疗的时代，每个与复杂疾病抗争的患者都应该基于分子层面的洞察，被引导至最佳治疗方案，”Illumina首席商务官埃弗雷特·坎宁安在一份声明中表示。“我们设想了一个世界，在这个世界中，全面的分子分析作为护理标准的一部分得以普及——不仅在癌症领域，还包括心脏病学、神经病学、免疫学以及所有其他疾病类别。”