NEW YORK – Fresh on the heels of a seed funding round, UK genomics startup OutSee is ready to put its unique approach to phenotype prediction to use in identifying therapeutic targets in partnership with pharmaceutical companies.

纽约——英国基因组学初创公司OutSee刚刚完成一轮种子融资，现在准备将其独特的表型预测方法应用于与制药公司合作识别治疗靶点。

The Cambridge-based company developed a computational method called Nomaly for simultaneously predicting phenotypes along with their possible underlying biological mechanisms.

总部位于剑桥的公司开发了一种名为Nomaly的计算方法，用于同时预测表型及其可能的潜在生物学机制。

Julian Gough, CEO and founder of OutSee, said that Nomaly, which was published in

朱利安·古奇（Julian Gough），OutSee的首席执行官兼创始人表示，Nomaly已经发布。

Nature Communications

自然通讯

in 2023, provides a complementary approach to more traditional genomic biomarker identification methods.

在 2023 年，提供了一种与更传统的基因组生物标志物鉴定方法相辅相成的途径。

Most genomic interrogation approaches, Gough explained, seek to identify correlations between genetic variants and phenotypic traits, generally by starting with a phenotype of interest and working back towards the underlying genetics, such as a group of researchers

大多数基因组审讯方法，Gough解释说，是寻求确定遗传变异和表型特征之间的相关性，通常是先从感兴趣的表型开始，再回溯到其背后的遗传学，例如一群研究人员所做的那样。

recently did

最近做了

to uncover genetic variants potentially contributing to infertility. While these correlation-based methods have proven effective in many cases, they often struggle to interpret rare and low-frequency functional variants due to challenges in statistically evaluating rare events in population genetics..

以揭示可能导致不孕的基因变异。尽管这些基于相关性的方法在许多情况下已被证明有效，但由于在群体遗传学中对罕见事件进行统计评估存在挑战，它们常常难以解释稀有和低频的功能性变异。

In contrast, Nomaly begins with the genomic data, analyzing sequences corresponding to the functional units –– also called domains –– of proteins to identify genetic outliers based on the structural and functional effects of variants found within functional units. The distance, or dissimilarity, between outliers and the rest of the cohort is calculated through an application of spectral clustering, a technique developed for identifying similarities in image segmentation analysis, that is now used for biological tasks such as extracting features from gene co-expression networks and for .

相比之下，Nomaly 从基因组数据入手，分析与蛋白质功能单元（也称为结构域）相对应的序列，根据功能单元内变异的结构和功能影响来识别遗传异常值。异常值与队列其余部分之间的距离或差异通过谱聚类技术计算得出，该技术最初为图像分割分析中的相似性识别而开发，如今已被用于生物任务，例如从基因共表达网络中提取特征等。

identifying and clustering cells

识别和聚类细胞

in single-cell sequencing datasets.

在单细胞测序数据集中。

'It basically compares all of those variants between all the people all at once and places everybody somewhere on the landscape,' Gough said. '[The] final score that you get from that will be how much of an outlier your original input genome is, relative to the background for one disease.'

“它基本上一次性比较了所有人之间的所有变异，并将每个人定位在某种程度的景观上，”高夫说。“你从中得到的最终分数将表明你的原始输入基因组相对于某种疾病的背景有多大程度的异常。”

Medical conditions experienced by the individuals representing these outliers can then be used to generate hypotheses about the mechanisms driving various diseases.

这些异常值所代表的个体所经历的医疗状况可以用来生成关于各种疾病驱动机制的假设。

'Knowing which functional unit a variant lies in gives us an estimate of whether it's going to have a big or a small impact,' Gough said. 'We now need to know what are the phenotypes and diseases that impact could affect.'

“知道一个变异位于哪个功能单元可以让我们估计它会产生大影响还是小影响，”高夫说。“我们现在需要知道可能受到影响的表型和疾病是什么。”

Phenotypes potentially impacted by the functional units of interest are identified via the

受兴趣功能单元潜在影响的表型通过以下方式识别：

dcGO

dcGO

(Domain-Centric Gene Ontology) database, after which the most likely hypotheses are experimentally checked and validated

（以域为中心的基因本体）数据库，然后通过实验验证和确认最有可能的假设。

in vitro

体外

, which can reveal information about the mechanisms driving disease.

，这可以揭示有关疾病驱动机制的信息。

Gough said that the application of spectral clustering found in Nomaly is another novelty of the system.

高夫表示，Nomaly 中发现的谱聚类应用是该系统的另一项新颖之处。

Gough explained that a key strength of this approach is that it complements correlation-based methods by enabling researchers to ask new types of questions.

高夫解释说，这种方法的一个关键优势是它通过使研究人员能够提出新的问题类型，从而补充了基于相关性的方法。

'Instead of starting with a question and building a cohort, sequencing them, and seeing if you can find something to correlate, and then seeing if we know enough about biology to answer it,' Gough said, 'we're kind of doing the reverse thing. We're saying, what are all the possible questions that you could ask for which an answer can be found in the data.'.

“我们没有从一个问题开始，建立一个队列，对他们进行排序，看看是否能找到一些相关性，然后再看我们是否对生物学有足够的了解来回答它，”高夫说，“我们正在做相反的事情。我们说的是，你能提出的所有可能的问题中，哪些问题的答案可以在数据中找到。”

This also enables researchers to more thoroughly interrogate a given dataset, something that Gough calls a challenge, in light of the uneven progress between data acquisition and analysis technologies.

这也使研究人员能够更彻底地查询给定的数据集，鉴于数据获取和分析技术之间的进展不均衡，高夫称其为一项挑战。

'Both the technology and the investment in sequencing human genomes has grown super-exponentially,' he said, 'But I would say that the analysis and interpretation methods have not grown at the same rate.'

“人类基因组测序的技术和投资都以超指数级增长，”他说道，“但我要说的是，分析和解读方法并没有以相同的速度发展。”

Another strength of the technique is that it can be applied to small datasets comprising only a few dozen individuals and still return reliable results.

该技术的另一个优点是可以应用于仅包含几十个个体的小数据集，并且仍然返回可靠的结果。

'Even [with] 30 to 50 people, we've found results reasonably regularly,' Gough said. 'Numbers like that … are too small to do any kind of correlation.'

“即使有30到50人，我们也能相当规律地找到结果，”高夫说。“这样的数字……太小了，无法进行任何相关性分析。”

Nomaly is currently patented in China, Japan, Korea, the UK, and the EU. The company also has a patent continuation-in-part application underway in the US, although it is unclear when this may be completed.

诺莫利目前在中国、日本、韩国、英国和欧盟拥有专利。该公司还在美国进行了一项专利延续申请，但目前尚不清楚何时能完成。

OutSee recently raised £1.8 million ($2.4 million) in seed funding, following an earlier £500,000 ($680,000) grant from Innovate UK, a national innovation agency.

OutSee 近期在获得英国国家创新机构 Innovate UK 早期提供的 50 万英镑（68 万美元）拨款后，又筹集了 180 万英镑（240 万美元）的种子资金。

Gough said that while the seed round remains open in hopes of reaching its £2.5 million funding target, the amount raised so far provides a cash runway through the end of next year and enables the company to transition from a setup phase of business development to executing on its strategy.

高夫表示，虽然种子轮融资仍处于开放状态，希望能达到250万英镑的融资目标，但迄今为止筹集到的金额为公司提供了延续到明年年底的现金跑道，并使得公司能够从业务发展的筹备阶段过渡到战略执行阶段。

OutSee hopes both to uncover new therapeutic targets from its own internal data and to partner with pharmaceutical and biotech companies on large cohort datasets.

OutSee希望既能从其内部数据中发现新的治疗靶点，又可以与制药和生物技术公司在大规模队列数据集上合作。

With respect to internal data, the company is now in the process of generating 10 candidate molecules targeting disorders of the central nervous system (CNS), as well as rare and metabolic diseases. Gough said that these represent the main focus areas of the company's in-house program.

关于内部数据，该公司目前正在生成针对中枢神经系统 (CNS) 疾病、罕见病和代谢性疾病的 10 种候选分子。高夫表示，这些代表了公司内部项目的主要关注领域。

Gough said that OutSee has already engaged in some small collaborations that grew 'organically' from interactions at conferences and other events. The company recently completed a project with Cambridge-based Astex Pharmaceuticals, involving therapeutic targets for CNS disorders, although Gough couldn't provide more specifics.

高夫表示，OutSee已经进行了一些小型合作，这些合作是通过在会议和其他活动中的互动“有机”发展而来的。该公司最近与剑桥的Astex制药公司完成了一个项目，涉及中枢神经系统疾病治疗靶点，但高夫无法提供更多细节。

.

。

'Now we're starting to go out and look for new partnerships,' he said.

“现在我们开始走出去寻找新的合作伙伴，”他说。