PEAKS 培訓(xùn)課:AI模型在高通量蛋白質(zhì)組學(xué)中的應(yīng)用

在過去的幾年中，大規(guī)模的蛋白質(zhì)組學(xué)研究使我們能夠開始在蛋白質(zhì)水平上了解表型。與人類基因組中大約 20,000 個(gè)蛋白質(zhì)編碼基因相比，人體內(nèi)有超過 100,000 種不同的蛋白質(zhì)，并且由于可變剪接以及轉(zhuǎn)錄后和翻譯后加工，還存在數(shù)十萬種 proteoform。多個(gè)課題研究表明蛋白質(zhì)形態(tài)比其相應(yīng)的蛋白質(zhì)更好地描述蛋白質(zhì)水平的生物學(xué)并且是更具體的分化指標(biāo)。人工智能（AI）模型的賦能帶來了高通量蛋白質(zhì)組學(xué)的革命。

本次為期兩天的 PEAKS 培訓(xùn)課程，將介紹 AI 模型在高通量蛋白質(zhì)組學(xué)中的應(yīng)用；展示如何針對多種互補(bǔ)的質(zhì)譜方法，結(jié)合數(shù)據(jù)庫搜索和從頭測序的數(shù)據(jù)分析，達(dá)到在proteoform水平上的深度覆蓋。無論是對于剛剛進(jìn)入AI賦能的蛋白質(zhì)組領(lǐng)域的新手，還是有經(jīng)驗(yàn)的大隊(duì)列樣本的蛋白質(zhì)組學(xué)研究，都非常適合參加這個(gè)培訓(xùn)。

---

# 基于深度學(xué)習(xí)的從頭測序與數(shù)據(jù)庫搜索：助力深度、高通量蛋白質(zhì)組學(xué)研究  Deep learning-based de novo sequencing and database search for in-depth and high-throughput proteomics

整合數(shù)據(jù)庫搜索與從頭測序技術(shù)，并結(jié)合多種酶解策略，可實(shí)現(xiàn)蛋白質(zhì)組的深度覆蓋。覆蓋度可達(dá)到1.5萬個(gè)蛋白組中的100萬條獨(dú)特肽段，平均序列覆蓋率約為80%。

Integrating database search and de novo sequencing and combining multiple enzyme digests facilitated deeper proteome coverage up to a million unique peptides from 15,000 protein groups, with a median sequence coverage of approximately 80%.

# 高可信度翻譯后修飾（PTM）與序列變異分析  Confident PTM and Sequence Variants Profiling

借助基于人工智能驅(qū)動(dòng)的評分模型，PEAKS在鑒定翻譯后修飾（PTM）與序列變異時(shí)可實(shí)現(xiàn)超高可信度，有效解決了在低豐度修飾或序列變異檢測中常見的靈敏度與準(zhǔn)確性問題。

Leveraging AI-powered scoring models, PEAKS achieves ultra-high confidence in identifying post-translational modifications (PTMs) and sequence variants, addressing the common issue of sensitivity and accuracy in low-abundance modification or variant detection.

#多組學(xué)驅(qū)動(dòng)的高準(zhǔn)確度、高靈敏度免疫肽組分析  Multi-omics enabled immunopeptidome with higher accuracy and sensitivity
將LC-MS/MS數(shù)據(jù)與下一代測序（NGS）結(jié)果相結(jié)合，PEAKS 可實(shí)現(xiàn)全面的免疫肽組分析， 突破單一技術(shù)手段的局限，確保精準(zhǔn)鑒定主要組織相容性復(fù)合體(MHC)結(jié)合肽，為癌癥免疫治療研究提供支持。

Integrating LC-MS/MS data with next-generation sequencing (NGS) results, PEAKS enables comprehensive immunopeptidome profiling—overcoming the limitation of single-technology approaches and ensuring accurate identification of MHC-bound peptides for cancer immunotherapy research.

#基于DIA的高效蛋白質(zhì)組學(xué)工作流程  Stream-lined Proteomics Proteomics Workflow with DIA
譜圖庫搜索與序列庫搜索是DIA數(shù)據(jù)分析的兩大核心方法。借助優(yōu)化后的從頭測序技術(shù)，PEAKS DIA 數(shù)據(jù)分析可實(shí)現(xiàn)蛋白質(zhì)組的深度覆蓋。
Spectral library search and library-free search are the two major approaches for DIA data analysis. Enhanced by de novo sequencing, PEAKS DIA data analysis provides you deep coverage.
#利用 PEAKS 從 DIA 數(shù)據(jù)中挖掘肽組與蛋白質(zhì)組“暗物質(zhì)”  Uncover dark peptidome/proteome from DIA data with PEAKS

除譜圖庫搜索與直接數(shù)據(jù)庫搜索外，我們提出一種新型計(jì)算方法，可直接從復(fù)雜 DIA圖譜中鑒定肽序列變異體與新肽段，同時(shí)嚴(yán)格控制假發(fā)現(xiàn)率（FDR）。

In addition to library search and direct database search, we present a novel computational method to identify peptide sequence variants and novel peptides directly and solely from complex DIA spectra while rigorously controlling the false discovery rate.

#結(jié)合Intact,Top-Down與Bottom-Up技術(shù)的深度Proteoform分析  Deep Proteoform Profiling with Intact, Top-Down and Bottom-Up

通過整合Intact、Top-Down(TD)，與Bottom-Up(BU)技術(shù)，PEAKS可實(shí)現(xiàn)全面的Proteoform表征（包括亞型、翻譯后修飾組合），這一突破解決了單一方法工具無法充分解析蛋白質(zhì)組復(fù)雜性的難題。

By integrating intact protein analysis, Top-Down (TD), and Bottom-Up (BU) approaches, PEAKS enables comprehensive proteoform characterization (including isoforms, PTM combinations)—a breakthrough for understanding proteome complexity that single-method tools cannot achieve.

#PEAKS GlycanFinder 實(shí)操演示 — 具有結(jié)構(gòu)分辨率的糖譜學(xué)分析  PEAKS GlycanFinder Walkthrough - Glycan Profiling with Structural
與只能檢測糖基是否存在的通用工具不同，GlycanFinder 可實(shí)現(xiàn)糖鏈結(jié)構(gòu)解析（如連接方式、分支結(jié)構(gòu)），解決了在糖蛋白組學(xué)研究中表征復(fù)雜糖鏈結(jié)構(gòu)的關(guān)鍵挑戰(zhàn)。
Unlike generic tools that only detect glycan presence, GlycanFinder provides structural-level resolution of glycans (e.g., linkage, branching), solving the key challenge of characterizing complex glycan structures in glycoproteomics research.

#基于云服務(wù)器的大規(guī)模隊(duì)列蛋白質(zhì)組學(xué)擴(kuò)展分析  Cloud-based scalable data analysis of large cohort proteomics.
PEAKS Online 云平臺(tái)可實(shí)現(xiàn)大規(guī)模隊(duì)列項(xiàng)目的自動(dòng)化分析，支持深度蛋白質(zhì)組學(xué)與蛋白質(zhì)基因組學(xué)分析，并能滿足發(fā)現(xiàn)型與靶向型質(zhì)譜技術(shù)的分析通量需求。

PEAKS Online enables the automation of large-scale cohort project with in-depth proteomics and proteogenomics analyses and throughput of discovery and targeted mass-spectrometry-based technologies.

#蛋白質(zhì)從頭測序  Protein de novo Sequencing 作為生物信息學(xué)的領(lǐng)軍企業(yè)，BSI專注于蛋白質(zhì)組學(xué)和生物藥領(lǐng)域，通過機(jī)器學(xué)習(xí)和先進(jìn)算法提供世界領(lǐng)先的質(zhì)譜數(shù)據(jù)分析軟件和蛋白質(zhì)組學(xué)服務(wù)解決方案，以推進(jìn)生物學(xué)研究和藥物發(fā)現(xiàn)。我們通過基于AI的計(jì)算方案，為您提供對蛋白質(zhì)組學(xué)、基因組學(xué)和醫(yī)學(xué)的卓越洞見。旗下著名的PEAKS®️系列軟件在全世界擁有數(shù)千家學(xué)術(shù)和工業(yè)用戶，包括：PEAKS®️ Studio，PEAKS®️ Online，PEAKS®️ GlycanFinder, PEAKS®️ AB，，ProteoformXTM ，DeepImmu®️免疫肽組發(fā)現(xiàn)服務(wù)和抗體綜合表征服務(wù)等。Email:  sales-china@bioinfor.com  電話：021-60919891

---