不久前，一家全球药企通过药明康德一体化平台开展了一系列多肽化合物相关项目。随着研发不断推进，一个现实问题摆在项目团队面前：面对结构日益复杂的多肽分子，如何更快、更准确地完成代谢产物的鉴定？
这并非简单的“加快实验进度”问题。
在部分复杂多肽项目中，传统数据处理和结构解析方式不仅耗时较长，且高度依赖研究人员的专业经验。当分子涉及环化、特殊修饰等复杂结构时，现有通用工具的适用范围也可能受到限制。
最终，药明康德药物代谢与动力学（DMPK）团队介入，借助自主开发的TidesID®一站式化合物鉴定平台，完成了相关复杂结构的解析，并在9个工作日交付项目结果。此后，该合作伙伴陆续将更多相关项目交由药明康德开展。
对DMPK团队而言，这个案例的意义远不止于“9个工作日”。它折射出新分子研发浪潮下日益突出的共性课题：当药物分子不断变化，与之配套的研究工具和工作模式，是否也需要同步革新？
图片来源：123RF
新分子正在改变代谢产物鉴定的“题目”
代谢产物鉴定（Metabolite Identification，MetID）是药物研发过程中的关键环节。
简单来说，研究人员需要从复杂的分析数据中寻找药物进入生物体系后的代谢产物，进一步解析其结构，并梳理药物可能的代谢途径。这些信息为早期分子结构优化、种属差异评估以及后续分析物选择等研发决策提供支撑。
在传统小分子领域，行业已经积累了相对成熟的研究方法和分析工具。然而，多肽、寡核苷酸等新分子的快速发展，正在改写这道“题目”。
与传统小分子相比，这些分子通常分子量更大、结构更复杂。一个分子可能包含不同的环化方式、骨架设计和化学修饰；随着组成单元增加，潜在代谢产物数量亦显著上升。分析仪器可以产生海量数据，但关键挑战在于如何从中提取有意义的信息。
传统模式下，这项工作很大程度上依赖研究人员逐一解析谱图，并结合专业经验进行判断。面对复杂分子，数据处理和结构解析时间随之延长。更重要的是，不同类型的新分子具备独特的结构特点，难以用一套固定流程解决所有问题。
商业化软件同样面临适配难题——一些高度复杂的分子结构甚至无法直接输入现有系统；而创新分子的设计正在快速变化，通用软件的功能迭代也未必能够及时覆盖新的项目需求。
这给DMPK团队带来现实挑战：随着复杂项目不断增加，仅靠扩充人力并非长久之策。
从几个小程序开始
TidesID®的起点，并非一项预先规划的大型软件开发项目。
最初，DMPK团队的一位科学家结合自身代码编写经验，以及对多肽代谢特点的理解，尝试开发了几个用于辅助数据处理的小程序。
出发点很简单：能否将那些重复、繁琐且遵循一定分析逻辑的工作交给程序处理？
实际项目验证后，这些小程序逐渐显现出价值。但随着项目数量增加，新需求不断涌现——多肽分子结构日趋复杂，寡核苷酸项目也快速增长。起初用于解决特定问题的小程序，开始面对越来越多不同的分子架构和分析场景。
DMPK团队由此开始与药明康德内部自动化技术团队协同，对原有工具进行系统性升级。
这是一个典型的跨学科合作过程。
DMPK科学家熟悉真实项目场景，他们清楚研究中可能遇到的特殊结构、最难处理的数据，以及分析流程中真正影响效率的环节。
自动化技术团队则汇集了结构解析、化学计量学、算法开发和软件工程等不同领域的专业人才，任务是将科学家长期积累的研究思路逐步转化为计算机可识别、计算和执行的分析逻辑。
以结构识别为例——看似只是软件流程的第一步，却直接关系到系统能否正确“理解”复杂分子结构并开展后续分析。
DMPK团队基于长期项目经验，梳理不同修饰方式和骨架设计所对应的结构类型；结构解析和化学计量学专家进一步建立计算逻辑；软件工程人员设计信息输入方式和系统架构，将不同模块整合进完整流程。
科学研究经验与数字化能力由此逐渐融合。
最初的几个“小程序”，逐步演变为TidesID®一站式化合物鉴定平台。目前，TidesID®已迭代至2.0版本，进一步拓展了多肽和寡核苷酸等复杂分子的代谢产物解析需求。
让复杂问题变得更可计算、更可复制
TidesID®试图解决的，不只是“让计算机跑得更快”。
传统模式下，复杂新分子的解谱高度依赖研究人员的专业积累。面对同一组数据，资深科学家与新进研究人员的处理效率和判断速度可能存在显著差异——这也是许多高度专业化领域共同面临的难题：如何让长期积累的专家经验得到更稳定的传承和应用？
在开发TidesID®过程中，DMPK团队逐步拆解长期项目实践中积累的分析思路，将复杂化合物鉴定问题转化为可计算、可执行的标准化分析流程。
系统首先识别分子结构，再根据结构特征建立相应分析逻辑，辅助研究人员开展数据解析。过去需要大量人工逐步排查的信息，可通过系统进行更有条理、更系统的处理。
人的作用并没有因此消失。
恰恰相反，研究人员可以将更多时间投入科学判断和复杂问题分析，而将部分重复性的数据处理工作交由自动化工具完成。
这种变化也使项目执行更加稳定和可控。
根据DMPK团队项目实践，应用TidesID®后，整体MetID项目周转时间可缩短约30%。对于结构更复杂、传统人工处理耗时更长的分子，自动化数据处理带来的效率提升尤为明显。在实际多肽项目中，复杂分子的解析周期已接近常规小分子项目的周转时间。文章开头提到的9个工作日交付案例，正是在这一背景下实现。
但效率并不是TidesID®的唯一目标。
对于代谢产物鉴定而言，数据完整性同样至关重要。TidesID®可覆盖不同类型的多肽结构，帮助研究人员更系统地识别潜在代谢产物，降低关键信息遗漏风险，使项目分析过程更加稳定、可控。
因此，对DMPK团队而言，开发TidesID®始终围绕一个核心目标：通过更系统的数据处理方式，为科学家提供更完整、更稳定的信息基础，从而更好地支持后续研发判断。
TidesID®背后：一场持续多年的自动化探索
如果只把TidesID®视为一款软件，或许难以完整理解其出现的背景。
药物研发中的自动化，并非简单地用机器取代人工操作。真正的难点在于：先理解一项科学流程为何如此运行，再判断哪些环节可标准化、哪些信息可数字化，以及如何将实验、数据与最终输出更顺畅地连接。
药明康德DMPK团队对自动化的探索已经持续多年——从早期引入自动化液体处理设备，到逐步推动独立实验环节的仪器化，再到进一步打通样品处理流程，自动化的边界持续向前延伸。
如今，在部分工作流程中，样品处理完成后可通过自动化物流系统进入分析环节；仪器完成数据采集后，系统进一步开展数据处理，并结合机器人流程自动化（RPA）支持后续报告流程。
团队正尝试将这些过去相对独立的环节逐步连接。
然而，在复杂新分子的代谢产物鉴定中，自动化所面对的问题有所不同。
硬件可以移动样品、完成操作，但真正决定解谱效率的，往往是数据背后的分析逻辑。
从这一角度看，TidesID®可理解为整个自动化体系中的“分析大脑”——它所承载的，是DMPK科学家在长期项目实践中积累的分析思路，以及对复杂分子结构和代谢行为的理解。
这也解释了为什么TidesID®需要由DMPK科学家与自动化专家共同开发。
缺少真实项目经验，算法难以准确理解待解决的问题；而没有软件与算法能力，科学家的专业经验也难以转化为可复制、可规模化应用的工具。
TidesID®的出现，正是长期科学积累与自动化能力不断融合的结果。
当分子不断变化，研究体系也需要持续进化
新分子研发仍在不断拓展药物创新的边界。新的结构、新的作用机制持续涌现，也不断提出过去未曾面对的研究问题，很难期待一套固定方法能回答所有问题。
这也是药明康德DMPK团队在新分子研究中长期坚持的思路：不存在适用于所有分子的“一刀切”方案。研究体系既需要理解不同分子的结构特点，也需要随着真实项目实践不断迭代。
从人工解谱，到商业化软件辅助，再到自主开发小程序和TidesID®自动化平台，这一过程并非源于对某项技术趋势的简单追逐。
很多时候，创新仅始于一个非常具体的问题：为什么一个项目耗时如此长？流程中的哪一步最依赖个人经验？哪些重复性工作可以交给计算机？当现有工具无法处理新的分子结构时，是否有其他解决路径？
不断提出问题，再从实际项目中寻找答案，最终推动了一套新的工作方式逐渐形成。
而TidesID®仍在继续迭代。未来，团队计划进一步探索将报告撰写、质量控制等后续环节纳入更完整的数字化工作流程，使数据从产生、分析到输出的连接更加顺畅。
CRDMO平台创新的缩影
对于药明康德而言，TidesID®也是一体化CRDMO平台持续支持新分子研发的缩影。
在CRDMO平台下，问题首先来自项目实践。科学家识别出共性挑战后，不同专业背景的团队可围绕同一个科学问题展开协作，将研究经验逐步转化为新方法、流程或工具；这些解决方案又在后续项目中持续验证和迭代，并进一步沉淀为平台能力。
TidesID®的发展正体现了这一过程。从DMPK科学家在复杂新分子项目中发现解谱效率和数据处理的挑战，到自动化、结构解析、算法和软件工程等专业能力参与其中，再到工具在真实项目中持续升级——其背后并非孤立的软件开发工作，而是一套由真实研发需求推动的能力迭代机制。
因此，TidesID®的意义并不仅限于缩短一次代谢产物鉴定的周期。它所体现的，是药明康德如何从真实研发场景中识别问题，依托跨学科协同形成解决方案，再通过持续的项目实践，将经验逐步沉淀为能够支持更多创新项目的平台能力。
而所有效率提升的最终意义，仍然指向同一个目标——让科学家能够更快获得可靠的信息，更有信心地做出下一步研发决策，帮助有潜力的创新分子更加稳步地向临床和患者迈进。
Complex Peptide Metabolite Identification in 9 Working Days: A WuXi AppTec’s DMPK Automation Case Study
Key Takeaways
Complex peptides can create significant challenges for metabolite identification because their higher molecular weights, diverse cyclization strategies, scaffold designs, and chemical modifications increase structural and analytical complexity.
TidesID® is WuXi AppTec’s proprietary automated spectrum interpretation platform for new modalities, supporting metabolite identification for complex molecules including peptides and oligonucleotides.
Based on WuXi AppTec’s DMPK project experience, TidesID® can shorten the overall turnaround time for metabolite identification (MetID) projects by approximately 30%; one complex peptide project was delivered within 9 working days.
Not long ago, a global pharmaceutical company was advancing a series of peptide-related projects through WuXi AppTec’s integrated platform. As development progressed, a new challenge emerged: as peptide structures become increasingly complex, how could metabolite identification be completed faster and with greater accuracy?
This was not simply a matter of “speeding up an experiment.”
For some complex peptides, traditional data processing and structural elucidation approaches can be time-consuming and highly dependent on the expertise of individual scientists. When molecules contain complex features such as cyclization or specialized chemical modifications, the applicability of commonly used analytical tools may also become limited.
WuXi AppTec’s Drug Metabolism and Pharmacokinetics (DMPK) team ultimately took on the challenge. Leveraging TidesID® intelligent one-top interpretation platform (TidesID®), a proprietary automated spectrum interpretation platform developed in-house, the team completed the relevant structural analysis and delivered the project results within 9 working days.
The client subsequently entrusted WuXi AppTec with additional related projects.
For the DMPK team, however, the significance of this case went well beyond “9 working days.” It reflected a broader challenge that is becoming increasingly visible as new modalities advance: as drug molecules evolve, must the research tools and workflows used to study them evolve as well?
Image Source: 123RF
Why Is Metabolite Identification More Challenging for Complex Peptides and New Modalities?
Metabolite identification, or MetID, is an important component of drug research and development. In simple terms, scientists examine complex analytical data to identify metabolites generated after a drug enters a biological system, further elucidate their structures, and map potential metabolic pathways. These insights can support decisions related to early-stage molecular optimization, cross-species metabolism assessment, and analyte selection for subsequent studies.
For traditional small molecules, the industry has accumulated relatively well-established research approaches and analytical tools. The rapid development of peptides, oligonucleotides, and other emerging modalities, however, is changing the nature of the challenge.
Compared with traditional small molecules, these molecules often have larger molecular weights and more complex structures. A single molecule may incorporate different cyclization strategies, scaffold designs, and chemical modifications. As the number of structural building blocks increases, the number of potential metabolites that may need to be considered can also rise substantially.
Analytical instruments can generate large volumes of data. The real challenge lies in identifying the information that matters.
Traditionally, this work has relied heavily on scientists reviewing spectra individually and applying their professional judgment. As molecular complexity increases, the time required for data processing and structural interpretation can increase accordingly. More importantly, different types of new modalities have distinct structural characteristics, making it difficult to address every molecule with a single standardized workflow.
Commercial software can face similar limitations. Some highly complex molecular structures may be difficult to input directly into existing systems. At the same time, as molecular designs continue to evolve rapidly, updates to general-purpose software may not always keep pace with emerging project needs.
For the DMPK team, this raised a practical question: as the number of complex projects continued to grow, simply adding more people would not provide a sustainable long-term solution.
How Was TidesID® Developed for Peptide and Oligonucleotide Metabolite Identification?
TidesID® did not begin as a large-scale software development initiative designed in advance.
Initially, a scientist on the DMPK team combined coding experience with an understanding of peptide metabolism to develop several small programs that could assist with data processing.
The idea was straightforward: could repetitive and time-consuming tasks that followed identifiable analytical logic be handled by software?
After being tested in real projects, these programs gradually demonstrated their value.
As project volumes increased, however, new requirements continued to emerge. Peptide structures became increasingly complex, while the number of oligonucleotide projects grew rapidly. Programs initially created to solve specific problems began to encounter a much broader range of molecular architectures and analytical scenarios.
The DMPK team therefore began collaborating with WuXi AppTec’s internal automation technology team to systematically upgrade the original tools.
It was a highly interdisciplinary process.
DMPK scientists understood the realities of project execution. They knew which unusual structures could arise, which data sets were the most difficult to interpret, and which steps in the analytical workflow had the greatest impact on efficiency.
The automation technology team, meanwhile, brought together expertise in structural interpretation, chemometrics, algorithm development, and software engineering. Their role was to translate the analytical expertise accumulated by scientists into processes that a computer could recognize, calculate, and execute.
Structure recognition is one example of this interdisciplinary collaboration. Although it is only the first step in the software workflow, it directly determines whether the system can correctly “understand” a complex molecular structure and perform subsequent analyses.
Drawing on extensive project experience, the DMPK team categorized structural types associated with different modifications and scaffold designs. Experts in structural interpretation and chemometrics then developed the computational logic, while software engineers designed the data input methods and system architecture needed to integrate individual modules into a complete workflow.
Scientific experience and digital capabilities gradually converged.
What began as several small programs ultimately evolved into TidesID®, a one-stop automated spectrum interpretation platform for new modalities. Today, TidesID® has advanced to version 2.0 and has expanded to support metabolite analysis needs for complex molecules including peptides and oligonucleotides.
How Does TidesID® Automate Spectrum Interpretation and Metabolite Identification?
TidesID® was not designed simply to “make computers run faster.”
Under traditional workflows, spectrum interpretation for complex new modalities can depend heavily on the accumulated expertise of individual scientists. When reviewing the same data set, a scientist with years of project experience and someone newly exposed to the molecule type may differ significantly in processing efficiency and speed of interpretation.
This reflects a broader challenge across highly specialized research fields: How can expert expertise be embedded in standardized workflows and applied more consistently?
During the development of TidesID®, the DMPK team progressively deconstructed analytical approaches accumulated through years of project experience and translated complex compound identification problems into standardized, computable, and executable analytical workflows.
The system first recognizes the molecular structure, then establishes an analytical framework based on its structural characteristics and assists scientists with data interpretation. Information that previously required extensive step-by-step manual review can be handled in a more structured and systematic manner.
The role of scientists does not disappear.
On the contrary, researchers can devote more time to scientific judgment and the analysis of complex questions, while automated tools handle part of the repetitive data-processing workload. This shift also helps make project execution more consistent and controllable.
Based on the DMPK team’s project experience, the use of TidesID® can shorten overall MetID project turnaround time by approximately 30%. For structurally more complex molecules that traditionally require longer periods of manual processing, the efficiency gains from automated data analysis can be particularly significant. In actual peptide projects, the analysis timeline for complex molecules can now approach the turnaround time typically seen in conventional small-molecule projects. The 9-working day delivery described at the beginning of this article was achieved in this context.
Efficiency, however, is not the only priority for TidesID®.
In metabolite identification, data completeness is equally important. TidesID® can support the analysis of different types of peptide structures, helping scientists identify potential metabolites more systematically, reduce the risk of missing critical information, and make project analysis more consistent and controllable.
For the DMPK team, the development of TidesID® has therefore remained focused on one central objective: to provide scientists with a more complete and consistent evidence base through more systematic data processing, thereby better supporting subsequent development decisions.
How Does Automation Support DMPK and Metabolite Identification Workflows?
Viewing TidesID® solely as a software tool may make it difficult to fully understand how it came into being.
Automation in drug development is not simply about replacing manual operations with machines. The greater challenge is first understanding why a scientific workflow functions the way it does, then determining which steps can be standardized, which information can be digitized, and how experiments, data, and final outputs can be connected more seamlessly.
WuXi AppTec’s DMPK team has been exploring automation for many years.
From introducing automated liquid-handling systems to automating individual experimental steps and connecting sample-processing workflows, the scope of automation has continued to expand.
Today, in selected workflows, samples can move from processing into analytical steps through automated logistics systems. Once instruments complete data acquisition, systems can perform further data processing and use robotic process automation, or RPA, to support downstream reporting activities.
The team is working to progressively connect steps that were previously relatively independent.
For metabolite identification of complex new modalities, however, automation presents a different type of challenge.
Hardware can transport samples and perform operations. But the factor that often determines the efficiency of spectrum interpretation is the analytical logic behind the data.
From this perspective, TidesID® can be viewed as an “analytical brain” within the broader automation system. It incorporates analytical approaches developed by DMPK scientists through long-term project experience, as well as their understanding of complex molecular structures and metabolic behavior.
This also explains why TidesID® required close collaboration between DMPK scientists and automation specialists.
Without real project experience, an algorithm may struggle to understand the problem it is expected to solve. Without software and algorithmic capabilities, scientific expertise can be difficult to translate into tools that are reproducible and scalable.
TidesID® emerged from the continued convergence of long-standing scientific expertise and automation capabilities.
Why Must Analytical Workflows Evolve for New Drug Modalities?
New modalities continue to expand the boundaries of drug innovation. New structures and mechanisms of action are emerging, bringing with them research questions that may not have existed before.
It is difficult to expect one fixed methodology to answer every question.
This reflects a principle the WuXi AppTec DMPK team has long applied in new modality research: there is no one-size-fits-all solution for every molecule. Research systems must account for the structural characteristics of different molecules and continue to evolve through real project experience.
The progression from manual spectrum interpretation to commercial software support, internally developed programs, and ultimately the TidesID® automated platform was not driven by the pursuit of a particular technology trend.
Innovation often begins with a very specific question: Why does a project take so long? Which step in the workflow is most dependent on individual expertise? Which repetitive tasks could be assigned to computers? When existing tools cannot process a new molecular structure, is there another way forward?
By continuing to ask these questions and searching for answers through real projects, a new way of working gradually took shape.
TidesID® itself continues to evolve. Looking ahead, the team plans to further explore the integration of report preparation, quality control, and other downstream activities into a more complete digital workflow, enabling smoother connections from data generation and analysis through final output.
How Can an Integrated CRDMO Platform Support New Modality Development?
For WuXi AppTec, TidesID® also reflects how an integrated CRDMO platform can continuously build capabilities in response to the evolving needs of new modality development.
Within WuXi AppTec’s CRDMO platform, capability development often begins with questions identified through project work. When scientists recognize a recurring challenge, teams with different areas of expertise can collaborate on the same scientific problem and progressively translate project experience into new methods, workflows, or tools. These solutions can then be tested, refined, and expanded through subsequent projects, gradually becoming broader platform capabilities.
The evolution of TidesID® illustrates this process. DMPK scientists first identified challenges related to spectrum interpretation efficiency and data processing while working on complex new modality projects. Specialists in automation, structural interpretation, algorithm development, and software engineering then worked together to address those challenges. As TidesID® continued to evolve through its use in real projects, it became more than an isolated software development initiative. It reflected a capability-building process driven by actual R&D needs.
The significance of TidesID® therefore extends beyond shortening the turnaround time for an individual metabolite identification project. It demonstrates how WuXi AppTec identifies challenges in real R&D settings, develops solutions through interdisciplinary collaboration, and translates accumulated project experience into platform capabilities that can support a broader range of innovative drug programs.
Ultimately, the purpose of greater efficiency remains the same: to help scientists obtain reliable information faster, make subsequent development decisions with greater confidence, and support promising new modalities as they advance toward clinical development and, ultimately, patient use.
免责声明：本文仅作信息交流之目的，文中观点不代表药明康德立场，亦不代表药明康德支持或反对文中观点。本文也不是治疗方案推荐。如需获得治疗方案指导，请前往正规医院就诊。
版权说明：欢迎个人转发至朋友圈，谢绝媒体或机构未经授权以任何形式转载至其他平台。转载授权请在「药明康德」微信公众号回复“转载”，获取转载须知。
分享，点赞，在看，聚焦全球生物医药健康创新
