3 研究方法 架設MVC資料庫
3.8.4 雙群落交集分析
兩個群落做交集分析時，因為兩個群落可能來自不統版本平台的晶片，
一般費雪檢定(Fisher exact test)(圖 二十二)(公式八)會因為兩群落母群體不同 而造成誤差，因此 AryNet 採用 Kupershmidt 於 2010 提出的雙基因群分析演算 法(公式九)，是一種從費雪檢定法再做改良的超幾何分布公式，用來計算兩個 集合的交集重疊是否具統計意義(圖 二十三)。
圖 二十二 費雪檢定法式意圖
利用超幾何分布公式計算 B2 和 B1 兩個集合交集是否具有統計意義。
𝑝(𝐵1, 𝐵2) = ∑ 𝐶𝐵1∩𝐵2𝐵1 𝐶𝐶𝑁−𝐵1∩𝐵2𝑁−𝐵1
𝐵2𝑁
𝐵2𝐵1∩𝐵2 (公式八)
圖 二十三 改良過費雪檢定
當兩個集合各自擁有不一樣的母集合時(例如兩組 dataset 分別來自不同平台)，
會因為過度估算總母集合而低估 p-value，修正過的公式只取中間兩平台交集來 做運算。
𝑝(𝐵1, 𝐵2) = ∑ 𝐶𝐵1∩𝐵2𝑃2∩𝐵1𝐶𝑃1∩𝑃2−𝐵1∩𝐵2𝑃1∩𝑃2−𝐵1
𝐶𝑃1∩𝐵2𝑃1∩𝑃2 𝐵2∩𝑃1
𝐵1∩𝐵2 (公式九)
環境賀爾蒙與精神疾病相關性分析 3.9
我們將 35 個化學物質(表 一)相關基因體實驗數據載入系統 ，顯著差異基 因(DEG, differential expression gene)過濾門檻設定為
1. T-test: p value 經過 FDR 校正後 q-value 需小於等於 0.05
部分腦切片的基因表現變異有高度重合(圖 二十六)。我們將「雙酚 A 樣本組基 因群落」中和「帕金森氏症黑質組織樣本組基因群落」重合的 132 個基因名稱 列出來並輸入 AryNet，其中表現量數據來源選擇「雙酚 A 10e-5M」，建立一 個指定基因名單群落(圖 二十七)。這個基因群落中共有 12 筆被 Rectome 資料 庫收入的蛋白質調控關係(圖 二十八)，其中尤以包括 ATR Serine/Threonine Kinase (ATR)的細胞分裂 S phase 的相關蛋白質反應途徑最多，另外我們將「基 因表現量相關度」節線顯示功能打開(圖 二十九)，將「雙酚 A 10e-5M」樣本 組中表現量相關度|r|>0.9 的基因連上虛線，然後找出網絡中中心度指標最高的 前 15 名基因，其中包括 PIK3R3、TUBB3、PSMD11 等(圖 三十)。
表 一 35種化學物質
Chemicals
Ampicillin CyclosporineA Genistein Propanil
Aresnic trioxide Daidzein Lindane Rapamycin
Azathioprine Deoxynivalenol Mannitol Silver nitrate
benzo[a]pyrene Diazinon Methylmercury Sodium citrate
Bis(2-ethylhexyl) phthalate Dibutyltin dichloride Mono-2-ethylhexyl phthalate Tributyltin chloride Bis(tributyltin) oxide Diethylstilbestrol Mycophenolic acid Urethane
Bisphenol A Fingolimod Nonylphenol 17alpha-ethynylestradiol
Cobalt2chloride Fluoxetine Ochratoxin A 17beta-estradiol
Cyclophosphamide Furosemide Prednisolone
表 二 35 種化學物質做 DEG 分析的 53 個基因群落內名單個數
Benzo[a]pyrene 0 0 0 17α-ethynylestradiol_dose1 1745 1903 3648 Bis(2-ethylhexyl) phthalate 0 1 1 17α-ethynylestradiol_dose3 0 0 0 Bis(tributyltin) oxide 0 0 0 17α-ethynylestradiol_dose2 1211 1114 2325
Cobalt2chloride 0 1 1 17β-estradiol_dose1 1127 1368 2495
表 三 4 種疾病做 DEG 分析的 19 個基因群落內名單個數 BA46,BA10為Boardman area編號
Disease ↑ ↓ Total
Bipolar disorder_dorsolateral prefrontal cortex 1 0 1 Bipolar disorder _dorsolateral prefrontal cortex_male 0 0 0 Bipolar disorder_dorsolateral prefrontal cortex_female 0 0 0 Bipolar disorder_orbital prefrontal cortex 0 1 1 Bipolar disorder_orbital prefrontal cortex_male 1 0 1 Bipolar disorder_orbital prefrontal cortex_female 0 0 0 Bipolar disorder_prefrontal cortex_BA46 0 0 0 Bipolar disorder_prefrontal cortex_BA10 0 0 0
Bipolar disorder_lymphocytes 0 0 0
Schizophrenia_prefrontal cortex_BA46 0 0 0 Schizophrenia_prefrontal cortex_BA10 0 0 0 Major depression_prefrontal cortex_BA10 0 0 0 Major depression_dorsolateral prefrontal cortex 0 0 0 Parkinson's disease_lateral substantia nigra 1595 1179 2774 Parkinson's disease_ lateral substantia nigra _male 1362 1140 2502 Parkinson's disease_ lateral substantia nigra _female 89 54 143 Parkinson's disease_prefrontal cortex_BA9 85 87 172 Parkinson's disease_prefrontal cortex_BA9_male 1 1 2 Parkinson's disease_prefrontal cortex_BA9_female 0 0 0
圖 二十四 化學物質與疾病基因群落交集分析過程圖
圖中最中間全部節點位置皆設為同一點的基因群落為躁鬱症的DEG群落，周邊 依序為Arsenictrioxide、Bisphathlate、Diazinon等11種化學物質的DEG群落，群 落間連線代表交集基因
圖 二十五 化學物質和精神疾病做雙群落交集分析結果
橫座標為8組精神病基因群落，縱座標為28組化學物基因群落，中間28x8=224 個色塊分別為所對應兩組基因群落交集分析後回傳的超幾何分布公式換算 p-value，在這色塊矩陣中因為p值<0.01而成鮮紅色的色塊共有31塊。
表 四 通過篩選的 31 對雙群落分析結果中，其中包含的化學物質 Diseases 欄位中 PD 為帕金森氏症， PFC 為額葉皮質，sng 為黑質。
化學物質中包含部分環境賀爾蒙、工業排放廢料、內分泌調節藥物、植物次級 代謝物等。
Diseases Chemical
PD_sng Arsenic trioxide, Lindane, Ochratoxin A, 17-alpha-ethynylestradiol, 17-beta-estradiol, BPA, Daidzein, Diethylstilbestrol, Genistein
PD_sng_male Arsenic trioxide, Lindane, Ochratoxin A, 17-alpha-ethynylestradiol, 17-beta-estradiol, BPA, Daidzein, Diethylstilbestrol, Genistein
PD_PFC_BA9 Lindane, Ochratoxin A, 17-alpha-ethynylestradiol, BPA, Diethylstilbestrol, Genistein
圖 二十六 雙酚A和帕金森氏症雙基因群落交集分析
其中左下角BPA_MCF10_10^5為雙酚A在10e-5M濃度下對MCF10細胞株造成的 顯著差異基因群落，另外三個為帕金森式症不同腦切片樣本組所統計出來的顯 著差異基因群落，群落間連線分別代表重合四個基因名單中彼此重合部分。紫 色線表示該基因在兩群落中「皆為正調控」(在兩個樣本組中皆為實驗組表現量 比對照組高)，橙色線表示「皆為負調控」(兩個樣本組中皆為對照組表現量較 高)，黑色線表示在兩群落中「調控方向相反」，其中可以看見三組帕金森氏症 相關基因群落彼此重合基因大多調控方向相同
圖 二十七 雙酚A群落與帕金森氏症群落交集的132個基因
在「BPA_MCF10_10^-5基因群落」中挑出和「Parkinson’s disease_lateral substansia nigra基因群落」重合的132個基因獨立建立一個新的基因群落，紅色、藍色分別 表示基因在「BPA_MCF10_10^-5基因群落」中的調控方向。其中可看到中間有 一群節點因為對應基因表現量彼此具有高度相關性(綠色、粉紅色虛線)而在圖 中集結一起。其中也有一些基因已經被Reactome收錄紀載彼此具有蛋白質交互 作用(實心箭頭、Z字線)。
圖 二十八 雙酚A群落與帕金森氏症群落交集的蛋白質交互作用
其中以ATR為主的6個相關基因皆為細胞分裂S-Phase生化途徑中的參與基因。其 他7筆蛋白質交互作用連線雖然尚無法找到明顯的共參與生化途徑，但仍能作 為後續研究的參考目標。
圖 二十九 132個重合基因於網絡中的中間度指標
「BPA_MCF10_10^-5基因群落」和「Parkinson’s disease_lateral substansia nigra基 因群落」重合的132個基因間開啟中間度指標功能後的圖表，以節點大小表示該 節點於網絡中的最短路徑中間度指標，其中節點越大表示該基因於此132個基因 組成網絡中越有可能扮演主要調控者的角色。
圖 三十132個重合基因中中間度指標最大的前15名基因
結果與討論
結論
未來研究方向 6
AryNet 目前僅提供人類基因體學相關資訊，且其收錄的生物晶片僅有 4 種 常用平台，未來希望能夠朝跨物種、多種平台來做擴充。為達此目的，必須安 裝更多生物晶片平台的相關演算法和晶片探針註解，並且需製作跨物種間同源 基因的比對檢索表，此功能若完成，將大大增加 AryNet 收錄資料庫以及統整 前人實驗的完整性。
另外 AryNet 內建的 REngine 計算引擎，目前尚無法進行平行運算，大幅 限制了 AryNet 在服務多使用者時的運算效能，也浪費伺服器本身的 CPU 資源。
目前市面上已有付費的 R 運算引擎-revolution R，即擁有平行運算能力，但其 程式撰寫技巧目前仍為商業機密，期盼未來能引用此技術以增進 AryNet 的統 計運算效能。
無論基因網絡的視覺化方式、使用者操作介面或是內部程式運作方法，
AryNet 都還有很大的改善空間，隨著硬體設備升級，AryNet 也將能提供更大 容量的計算。因此本研究系統以 MVC 方式分割撰寫，並且選用物件導向程式 設計，使其在日後維護與更新上都能夠有較大的彈性，期盼 AryNet 能在未來 基因體學以及病理因子研究領域中帶來正向幫助。
參考文獻 7
1. Tondo, L., et al., Suicide attempts in bipolar disorders: comprehensive review of 101 reports. Acta Psychiatr Scand, 2015.
2. Gabbard, G.O., Psychotherapy in psychiatry. Int Rev Psychiatry, 2007. 19(1): p. 5-12.
3. Lee, S.H., et al., Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs. Nat Genet, 2013. 45(9): p. 984-94.
4. Moutoussis, M., G.W. Story, and R.J. Dolan, The computational psychiatry of reward:
broken brains or misguided minds? Front Psychol, 2015. 6: p. 1445.
5. Etain, B., et al., Beyond genetics: childhood affective trauma in bipolar disorder. Bipolar Disord, 2008. 10(8): p. 867-76.
6. Maniglio, R., The impact of child sexual abuse on the course of bipolar disorder: a systematic review. Bipolar Disord, 2013. 15(4): p. 341-58.
7. Zhao, H., et al., Genome-wide DNA methylome reveals the dysfunction of intronic microRNAs in major psychosis. BMC Med Genomics, 2015. 8(1): p. 62.
8. Dong, E., et al., DNA-methyltransferase1 (DNMT1) binding to CpG rich GABAergic and BDNF promoters is increased in the brain of schizophrenia and bipolar disorder patients.
Schizophr Res, 2015. 167(1-3): p. 35-41.
9. Belvederi Murri, M., et al., The HPA axis in bipolar disorder: Systematic review and meta-analysis. Psychoneuroendocrinology, 2015. 63: p. 327-342.
10. Nestler, E.J., et al., Epigenetic Basis of Mental Illness. Neuroscientist, 2015.
11. Xu, Z., et al., DNA-based hybridization chain reaction amplification for assaying the effect of environmental phenolic hormone on DNA methyltransferase activity. Anal Chim Acta, 2014. 829: p. 9-14.
12. Tang, W.Y., et al., Neonatal exposure to estradiol/bisphenol A alters promoter methylation and expression of Nsbp1 and Hpcal1 genes and transcriptional programs of Dnmt3a/b and Mbd2/4 in the rat prostate gland throughout life. Endocrinology, 2012. 153(1): p. 42-55.
13. Kitraki, E., et al., Developmental exposure to bisphenol A alters expression and DNA methylation of Fkbp5, an important regulator of the stress response. Mol Cell Endocrinol, 2015. 417: p. 191-9.
14. Kundakovic, M., et al., DNA methylation of BDNF as a biomarker of early-life adversity.
Proc Natl Acad Sci U S A, 2015. 112(22): p. 6807-13.
15. Hovenkamp-Hermelink, J.H., et al., Low stability of diagnostic classifications of anxiety disorders over time: A six-year follow-up of the NESDA study. J Affect Disord, 2015. 190: p.
310-315.
16. Choudhuri, S., From Waddington's epigenetic landscape to small noncoding RNA: some important milestones in the history of epigenetics research. Toxicol Mech Methods, 2011.
21(4): p. 252-74.
17. Holliday, R., Epigenetics: an overview. Dev Genet, 1994. 15(6): p. 453-7.
18. Pandian, G.N. and H. Sugiyama, Strategies to modulate heritable epigenetic defects in cellular machinery: lessons from nature. Pharmaceuticals (Basel), 2012. 6(1): p. 1-24.
19. Vaiserman, A.M., Epigenetic programming by early-life stress: Evidence from human populations. Dev Dyn, 2014.
20. Liyanage, V.R., et al., DNA modifications: function and applications in normal and disease States. Biology (Basel), 2014. 3(4): p. 670-723.
21. Szyf, M., Lamarck revisited: epigenetic inheritance of ancestral odor fear conditioning. Nat Neurosci, 2014. 17(1): p. 2-4.
22. Mao, Z., et al., Paternal BPA exposure in early life alters Igf2 epigenetic status in sperm and induces pancreatic impairment in rat offspring. Toxicol Lett, 2015. 238(3): p. 30-8.
23. Pierre, J.M., Faith or delusion? At the crossroads of religion and psychosis. J Psychiatr Pract, 2001. 7(3): p. 163-72.
24. Belusa, L., A.M. Selzer, and B.D. Partecke, [Description of Dupuytren disease by the Basel physician and anatomist Felix Plater in 1614]. Handchir Mikrochir Plast Chir, 1995. 27(5): p.
272-5.
25. Jorgensen, K.N., et al., First- and second-generation antipsychotic drug treatment and subcortical brain morphology in schizophrenia. Eur Arch Psychiatry Clin Neurosci, 2015.
26. Edlinger, M., et al., Risk of violence of inpatients with severe mental illness--do patients with schizophrenia pose harm to others? Psychiatry Res, 2014. 219(3): p. 450-6.
27. Shibre, T., et al., Suicide and suicide attempts in people with severe mental disorders in Butajira, Ethiopia: 10 year follow-up of a population-based cohort. BMC Psychiatry, 2014.
14: p. 150.
28. Aydin, A., et al., Mood and metabolic consequences of sleep deprivation as a potential endophenotype' in bipolar disorder. J Affect Disord, 2013. 150(2): p. 284-94.
29. Song, J., et al., Bipolar disorder and its relation to major psychiatric disorders: a family-based study in the Swedish population. Bipolar Disord, 2015. 17(2): p. 184-93.
30. Szczepankiewicz, A., Evidence for single nucleotide polymorphisms and their association with bipolar disorder. Neuropsychiatr Dis Treat, 2013. 9: p. 1573-82.
31. Etain, B., et al., Childhood trauma is associated with severe clinical characteristics of bipolar disorders. J Clin Psychiatry, 2013. 74(10): p. 991-8.
32. Bratlien, U., et al., Environmental factors during adolescence associated with later development of psychotic disorders - a nested case-control study. Psychiatry Res, 2014.
215(3): p. 579-85.
33. Korach, K.S., M. Metzler, and J.A. McLachlan, Diethylstilbestrol metabolites and analogs.
New probes for the study of hormone action. J Biol Chem, 1979. 254(18): p. 8963-8.
34. Iguchi, T., [Environmental endocrine disruptors]. Nihon Rinsho, 1998. 56(11): p. 2953-62.
35. Bazer, F.W., et al., Environmental factors affecting pregnancy: Endocrine disrupters, nutrients and metabolic pathways. Mol Cell Endocrinol, 2014.
36. Kucka, M., et al., Atrazine acts as an endocrine disrupter by inhibiting cAMP-specific phosphodiesterase-4. Toxicol Appl Pharmacol, 2012. 265(1): p. 19-26.
37. Li, R., et al., Effects of DEHP on endometrial receptivity and embryo implantation in pregnant mice. J Hazard Mater, 2012. 241-242: p. 231-40.
38. Multigner, L., et al., [Environment and secular sperm trend. Stallion's semen quality during the last two decades]. Rev Epidemiol Sante Publique, 2000. 48 Suppl 2: p. 2s72-8.
39. Xu, H., et al., The impact of endocrine-disrupting chemicals on oxidative stress and innate immune response in zebrafish embryos. Environ Toxicol Chem, 2013. 32(8): p. 1793-9.
40. Perera, F., et al., Prenatal bisphenol a exposure and child behavior in an inner-city cohort.
Environ Health Perspect, 2012. 120(8): p. 1190-4.
41. Singh, S. and S.S. Li, Epigenetic effects of environmental chemicals bisphenol A and phthalates. Int J Mol Sci, 2012. 13(8): p. 10143-53.
42. Elyakim, E., et al., hsa-miR-191 is a candidate oncogene target for hepatocellular carcinoma therapy. Cancer Res, 2010. 70(20): p. 8077-87.
43. Kundakovic, M., et al., Sex-specific epigenetic disruption and behavioral changes following low-dose in utero bisphenol A exposure. Proc Natl Acad Sci U S A, 2013. 110(24): p.
9956-61.
44. Kruppa, J. and K. Jung, Set-Based Test Procedures for the Functional Analysis of Protein Lists from Differential Analysis. Methods Mol Biol, 2016. 1362: p. 143-56.
45. Mahajan, G. and S.C. Mande, From System-Wide Differential Gene Expression to Perturbed Regulatory Factors: A Combinatorial Approach. PLoS One, 2015. 10(11): p. e0142147.
46. Ogata, H., et al., KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res, 1999. 27(1): p. 29-34.
47. Vastrik, I., et al., Reactome: a knowledge base of biologic pathways and processes.
Genome Biol, 2007. 8(3): p. R39.
48. Mlecnik, B., et al., PathwayExplorer: web service for visualizing high-throughput expression data on biological pathways. Nucleic Acids Res, 2005. 33(Web Server issue): p.
W633-7.
49. Kutmon, M. and A. Riutta, WikiPathways: capturing the full diversity of pathway knowledge. 2015.
50. Viti, C., et al., High-throughput phenomics. Methods Mol Biol, 2015. 1231: p. 99-123.
51. Delgado, S., et al., A Novel Representation of Genomic Sequences for taxonomic clustering and visualization by means of Self-Organizing Maps. Bioinformatics, 2014.
52. Zheng, Y., et al., Standardization efforts enabling next-generation sequencing and microarray based biomarkers for precision medicine. Biomark Med, 2015.
53. Lim, W.K., et al., Comparative analysis of microarray normalization procedures: effects on
reverse engineering gene networks. Bioinformatics, 2007. 23(13): p. i282-8.
54. Millenaar, F.F., et al., How to decide? Different methods of calculating gene expression from short oligonucleotide array data will give different results. BMC Bioinformatics, 2006.
7: p. 137.
55. Zhao, X., et al., [Whole genome methylation profiles of myelodysplastic syndrome and its diagnostic value]. Zhonghua Xue Ye Xue Za Zhi, 2014. 35(10): p. 944-8.
56. Barrett, T., et al., NCBI GEO: mining millions of expression profiles--database and tools.
Nucleic Acids Res, 2005. 33(Database issue): p. D562-6.
57. Kurscheid, S., et al., Chromosome 7 gain and DNA hypermethylation at the HOXA10 locus are associated with expression of a stem cell related HOX-signature in glioblastoma.
Genome Biol, 2015. 16: p. 16.
Genome Biol, 2015. 16: p. 16.