淡水長臂大蝦之轉錄體組裝與多維體資料分析

Abstract

淡水長臂大蝦 (Macrobrachium rosenbergii) 為重要的淡水養殖蝦種，其產值在全球逐年增長，然而目前缺乏較完整的基因體與轉錄體資料，無法以更全面的方式來進行研究與智慧育種。因此本篇研究將組裝淡水長臂大蝦之轉錄基因體，利用NCBI Sequence Read Archive (SRA) 中淡水長臂大蝦 (Macrobrachium rosenbergii) 已發表文獻且定序資料有上傳之轉錄資料作為基礎，透過從頭組裝 (de novo assembly) 的策略，來建構完整轉錄基因體 (Transcriptome)。所使用的轉錄片段涵蓋後期蝦苗與其感染泰國蝦諾達病毒 (Macrobrachium rosenbergii nodavirus, MrNV) 的個體、稚蝦與其受哈維弧菌 (Vibrio harveyi) 感染的個體、成蝦與其感染白點症病毒 (Whispovirus) 的個體，共整合20筆轉錄體資料。初步組裝結果顯示，共有26.3萬多條轉錄序列，去除重複性達95%以上的剩餘20.9萬條，同時預測長度大於30個胺基酸的蛋白質有154,901筆資料，利用泛用單拷貝直系同源序列基準 (Benchmarking Universal Single-Copy Orthologs, BUSCO)，可以評估組裝體的完整度達到96.5% (arthopoda_odb10)。針對預測到之轉譯蛋白進行功能性註解，透過SignalP程式得到6,325條訊號肽、TMHMM程式得到11,944條穿膜蛋白、在Pfam資料庫中一共註解到199,898筆Pfam ID，也就是蛋白質的功能區塊，並且可以轉換出103,324筆GO項目、在NR資料庫中可註解共154,497筆資料，其中對應到數量最多的物種為美洲螯龍蝦 (Homarus americanus) 一共12,069條、KEGG代謝路徑一共可對應到424筆共11,609條。在新型抗菌肽 (antimicrobial peptides) 的鑑別部分，藉著人工智慧模型所開發之預測平台：AI4AMP進行判定，共得到7,913筆結果，再將此結果於無脊椎動物的AMP資料庫：InverPep中進行比對，找出一共13筆抗菌肽序列，包含crustin、Histone H2A、CgUbiquitin、kazal2、Techylectin-5B、MrH4，若將7,913筆與訊號肽的結果進行比對，共可以找出232筆潛在的抗菌肽序列。 為了瞭解與其他常見養殖蝦種之間的基因異同，本研究以四個屬於對蝦科 (Penaeidae) 的物種分別是白蝦 (Litopenaeus vannamei)、草蝦 (Penaeus monodon)、日本對蝦 (Marsupenaeus japonicus) 與中國對蝦 (Fenneropenaeus chinensis)、一個長臂蝦科 (Palaemonidae) 的物種為日本沼蝦 (Macrobrachium nipponense)，以及作為外群的端足蟲 (Hyalella azteca)，來進行比較基因體學的分析。透過親緣關係的解析，可以估計物種分化大概年代，結果顯示淡水長臂大蝦與日本沼蝦的共同祖先出現在約94百萬年前。藉由資料庫註解結果找出可能的甲殼類高血糖素 (crustacean hyperglycemic hormone, CHH) 家族一共8筆，並發現到淡水長臂大蝦與日本沼蝦CHHs的數量與多樣性都少於另外四種對蝦科，儘管日本沼蝦在109百萬年前發生全基因體複製 (whole genome duplication)，但CHHs數量並沒有隨著增加。比對8筆CHHs在NCBI中淡水長臂大蝦的序列，發現僅有3筆能完全對應，顯示淡水長臂大蝦的CHHs尚未完整地被建立。 本研究組裝了淡水長臂大蝦之轉錄基因體，並以此為出發點預測蛋白質序列，但缺乏DNA序列的定序資料，使基因體的完整度仍不及白蝦、草蝦等，若未來有更完整的序列，將有機會拼裝出染色體層級的基因體。最後，本研究整理了淡水長臂大蝦的基因體、轉錄體、蛋白質體與功能性註解結果，建構為線上資料庫：Multi-Omics online Analysis System (MOLAS)，並且額外新增淡水長臂大蝦各組織的RNA序列，以便進行不同組別間基因差異表現量的研究，以同樣的平台，我們也整理了白蝦、草蝦、中國對蝦、日本對蝦與日本沼蝦的資料，建立各自的線上平台，未來將提供東南亞常見蝦種的研究者，一個更友善的圖像化線上分析環境。

The giant river prawn (Macrobrachium rosenbergii) plays a pivotal role in aquaculture, a billion-dollar industry which is also in a growing demand to date. This study gather giant river prawn short reads in RNA-seq from published and NCBI Sequence Read Archive (SRA) to assemble a better-quality transcriptome. A total of 20 samples for sequencing are obtained from different stages of the life cycle such as post-larvae (PL), juvenile and adult, and conditions of infected with Macrobrachium rosenbergii nodavirus (MrNV), Vibrio harveyi, and Whispovirus (White Spot Syndrome Virus, WSSV). The preliminary assembly result is about 263 thousand contigs and next are clustered to eliminate those similarities reached 95%. About 209 thousand contigs are remained and according to the assessment of Benchmarking Universal Single-Copy Orthologs (BUSCO) program, the completeness was 96.5% (arthopoda_odb10). Then, those contigs are implemented in prediction and annotation procedures. With the limitation of the 30 amino acids threshold, 154,901 protein sequences are predicted. Through functional annotation, we obtain 6,325 signal peptides, 11,944 sequences contain transmembrane helices, 199,898 Pfam IDs which can convert to 103,324 GO terms, 154,497 NR database alignment results with the Homarus americanus comprising of 12,069 hits, 424 KEGG pathways, 13 AMPs (antimicrobial peptides) validated from InverPep database include crustin, Histone H2A, CgUbiquitin, kazal2, Techylectin-5B, MrH4 and 232 potentially AMPs which contain signal peptides. Other 6 species include Litopenaeus vannamei, Penaeus monodon, Fenneropenaeus chinensis, Marsupenaeus japonicus from Penaeidae, Macrobrachium nipponense from Palaemonidae and Hyalella azteca as outgroup are chosen to do the comparative research. The divergent time are inferred by single copy orthologues within 7 species. The common ancestor of M. nipponense and M. rosenbergii appeared in 94 million years ago. Further study focuses on crustacean hyperglycemic hormone (CHH), which comprise several structure-related but functionally diverse proteins. It is found that shrimps belonging to Penaeidae have a greater amount and greater diversity of CHHs than Palaemonidae. Although M. rosenbergii and M. nipponense are seemed to experience whole-genome duplication, their amount of CHH-related proteins is less than the other 4 Penaeidae species. There are 8 CHHs found in M. rosenbergii and only 3 of them can completely align to sequence of M. rosenbergii in NCBI database which indicate that CHHs of M. rosenbergii have not been established thoroughly. This study assembles the transcriptome of M. rosenbergii and predicts its protein sequence but due to the lack of DNA sequence, the genome of M. rosenbergii is not as complete as L. vannamei or P. monodon. In the end, all of the M. rosenbergii’s prediction and functional annotation data are stored into a web database, Multi-Omics online Analysis System (MOLAS), that are accessible to the public. Additional expression data from each tissue of M. rosenbergii are added to enrich comparison study. Species included L. vannamei, P. monodon, F. chinensis, M. japonicus and M. nipponense also have their own version of MOLAS. Others interested in these Asian common shrimps are welcome to use this user-friendly platform to do the research.