陳洛南研究組發(fā)表“由單樣本動態(tài)網(wǎng)絡(luò)標(biāo)志物檢測生物過程/疾病過程的臨界狀態(tài)及其關(guān)鍵分子”的研究成果
來源：
時間：2019-02-02
2018年12月28日，國際學(xué)術(shù)期刊National Science Review在線發(fā)表了中國科學(xué)院生物化學(xué)與細胞生物學(xué)研究所陳洛南研究組題為“Detection for disease tipping points by landscape dynamic network biomarkers”的最新研究成果。該成果首次建立單樣本“l(fā)andscape”動態(tài)網(wǎng)絡(luò)標(biāo)志物(DNB: dynamic network biomarker)理論和方法，實現(xiàn)基于單個樣本數(shù)據(jù)可檢測生物動態(tài)過程/疾病過程的臨界狀態(tài)及其關(guān)鍵分子。該理論結(jié)合單樣本網(wǎng)絡(luò)構(gòu)建方法和動態(tài)網(wǎng)絡(luò)標(biāo)志物理論，提出針對單樣本“l(fā)andscape”的動態(tài)網(wǎng)絡(luò)標(biāo)志物檢測方法，并利用該方法對癌癥數(shù)據(jù)進行了分析，不僅成功地檢測到疾病臨界狀態(tài)而且得到了在三種癌癥的DNB和“預(yù)后生物標(biāo)志物”。
在復(fù)雜疾病的研究中，疾病的前期預(yù)警或惡化的早期預(yù)警信號是復(fù)雜疾病預(yù)防和治療重要的診斷指標(biāo)，如果能夠成功的量化疾病臨界狀態(tài)或捕獲疾病的早期預(yù)警信號，對復(fù)雜疾病的預(yù)防和治療有著深遠意義。
基于這個問題，研究人員利用動態(tài)網(wǎng)絡(luò)標(biāo)志物的概念并結(jié)合單樣本網(wǎng)絡(luò)構(gòu)建方法，提出了在單樣本水平上，對每個基因/分子進行進行系統(tǒng)性甄別的“l(fā)andscape”方法。該方法可以對單次采樣的表達譜數(shù)據(jù)進行分析，有效地找到樣本特異性的動態(tài)網(wǎng)絡(luò)標(biāo)志物，并檢測其生物動態(tài)過程的臨界狀態(tài)或疾病的早期預(yù)警信號。通過對癌癥數(shù)據(jù)的分析，研究人員利用該方法成功找到了LUAD、THCA和KIRC三種癌癥的DNB和“預(yù)后生物標(biāo)志物”。該成果也可應(yīng)用于進化等復(fù)雜非線性生物過程的研究。
生化與細胞所陳洛南研究員與東京大學(xué)合原一幸教授為該本文的共同通訊作者，山東大學(xué)劉小平研究員和安徽財經(jīng)大學(xué)的常嘯副教授為本文的共同第一作者，該研究得到了科技部、中國科學(xué)院和國家自然科學(xué)基金的經(jīng)費支持。
(A) Schematic diagram for disease progression of a complex disease in a subject. There are three states during disease progression comprising a normal state, a critical state (or pre-disease state or tipping point), and the final disease state. Generally, the phenotypic and molecular expressions of the disease state are significantly different from those of the normal state, but there is no significant differences observed between the critical and the normal states. Thus, detection of the critical state is difficult. However, there are strong collective fluctuations in processes at the critical state, which differs from the other states. (B) The l-DNB flowchart for identifying DNBs from a single sample. The individual data of sample d is used to construct the SSN for sample d. For every gene x, its local module is comprised of gene x and its first-order neighbors. The local module score for gene x can indicate the local DNB score for the gene. After ranking the scores for all genes, the top-k genes can be regarded as the potential DNB for the sample d. The sEDn, sPCCin and sPCCout values for each individual sample are defined in the Methods.
