Maize B104 (beta) genome assembly and annotation

In the interest of the maize research community, the Iowa State Transformation Facility and Lawrence-Dill Plant Informatics and Computation Lab in collaboration with MaizeGDB has released a beta-version of the B104 maize genome and structural annotations prior to scientific publication in accordance with guidelines set forth by the Toronto agreement (http://www.nature.com/nature/journal/v461/n7261/full/461168a.html) for prepublication data sharing (Nature. 2009 461:168). The above groups reserve the first right to publish on the available B104 data including but not limited to whole-genome comparisons, genes, structural annotations, functional annotations, and genome-wide association studies. The group also reserves the right to the first opportunity to improve this sequence and its annotations for a full official genome release (version 1; anticipated release December 2017). Under the Toronto agreement, researchers can use the B104 sequence and annotation to study individual or small sets of genes and localized regions of the genome. Any redistribution of these data should include the full text of the data use policy.<br><br><br>Assembly Note: <br>The B104 genome was sequenced using Illumina paired end sequencing on HiSeq2000 at raw read coverage of 50X with read length of 101 bp and insert size 250 bp. Trimmed reads were assembled using a reference- guided assembly approach with B73 reference sequence (AGPv2 5b pseudo molecules). The reads were mapped to the B73 reference, consensus sequences were generated and gaps were filled. All paired end reads were mapped back to the gap-filled scaffolds to identify the regions with abnormal links and corresponding scaffolds were broken from those regions. Unmapped reads were assembled using a de novo approach and scaffolds were added to the final assembly.<br><br>B104 Gene Models:<br><br>B104 (beta) gene models were developed by the Lawrence-Dill Plant Informatics Lab in collaboration with MaizeGDB and Dow AgroSciences, and made available through www.maizegdb.org.<br><br>Genes were annotated using MAKER-P Pipeline, using both evidence-based approach (cDNA and EST data) and an ab initio approach. For evidence based prediction, transcriptome assemblies of diverse tissue types from B104 and other maize lines (B73, Miami white, Mo17, OH43, B97, W22, A188), full length cDNA sequences and annotated B73 proteins from NCBI, Arabidopsis thaliana, Oryza sativa japonica and other plant proteins from Uniprot were used. For ab initio predictions, AUGUSTUS and SNAP gene predictors were used. Gene models are named as recommended by the Maize Nomenclature Committee. <br>

为服务玉米研究共同体，爱荷华州立大学转化设施（Iowa State Transformation Facility）与劳伦斯-迪尔植物信息与计算实验室（Lawrence-Dill Plant Informatics and Computation Lab）联合玉米基因组数据库（MaizeGDB），遵照《多伦多协定》（Toronto agreement，http://www.nature.com/nature/journal/v461/n7261/full/461168a.html）中关于预发布数据共享的规范（《自然》*Nature*. 2009 461:168），在正式科学发表前发布了B104玉米基因组及其结构注释的测试版（beta-version）。

上述团队保留针对该B104数据的首次发表权，涵盖但不限于全基因组比较、基因、结构注释、功能注释及全基因组关联研究等方向。该团队同时保留对该序列及其注释进行优化，以推出正式完整基因组版本（版本1，预计2017年12月发布）的优先权利。根据《多伦多协定》，研究人员可使用B104序列及注释开展单个或少量基因、基因组局部区域的研究。若对这些数据进行任何再分发，需附带完整的数据使用政策文本。

组装说明：
B104基因组采用Illumina双端测序技术，在HiSeq2000测序平台上完成测序，原始读段覆盖度为50X，读长101 bp，插入片段长度250 bp。经修剪的读段采用参考引导组装策略，以B73参考序列（AGPv2 5b假分子）为参照进行组装。首先将读段比对至B73参考序列，生成共识序列并填补缺口。将所有双端读段回比至填补缺口后的支架序列，以识别存在异常连接的区域，并将对应支架从该位置断开。未比对上的读段采用从头组装（de novo）策略进行组装，所得支架添加至最终组装结果中。

B104基因模型：

B104（测试版）基因模型由劳伦斯-迪尔植物信息实验室联合MaizeGDB与陶氏益农（Dow AgroSciences）开发，并通过www.maizegdb.org对外发布。

基因注释采用MAKER-P流程（MAKER-P Pipeline），结合基于证据的方法（cDNA与EST数据）及从头预测（ab initio）方法开展。基于证据的预测使用了B104及其他玉米品系（B73、Miami white、Mo17、OH43、B97、W22、A188）的多样组织转录组组装结果、全长cDNA序列，以及来自美国国家生物技术信息中心（NCBI）的注释B73蛋白序列、拟南芥（Arabidopsis thaliana）、粳稻（Oryza sativa japonica）蛋白序列和UniProt数据库中的其他植物蛋白序列。从头预测则使用了AUGUSTUS与SNAP基因预测工具。基因模型的命名遵循玉米命名委员会（Maize Nomenclature Committee）的建议。