如何确定平台上的序列数据的可靠性-DarwinTree.DOCVIP

20091028反馈响应稿 V4 文本交流记 事项 时间 文稿版本 备注 反馈提交 2009-10-28 Version 1 仙湖会议 初次响应 2009-10-29 V2 孟珍 第二次响应 2009-10-29 V3 董慧 会议讨论 2009-10-30 V4 三方会议 palpp@ 2009-10 仙湖2009-10-23会议反馈 如何确保平台序列数据的可靠性和正确性。由于现行的序列清洗标准尚存在不足，导致在平台上检出的序列存在较多误差。因此，需要制定新的序列清洗标准，以确保检出序列的正确性和可靠性。这是保证平台质量的前提。 （1）关于序列过滤甄别原则，是否请红梅和老师同学们把规则细化？我们来程序实现？ （2）现在平台只是普遍性的过滤，使用具体问题和问题归结研究，请红梅和老师同学们共同指导完成。2009-10-28数据有加了过滤规则，已进行新的数据清洗。请大家再使用反馈！ （3）信息团队【Inf】内部，意在研究考虑blast抽提应用。 此部分需要大家多商定确认。希望双方共同合作推进！ 关于序列的甄别，其他的平台上这样的一些方法。 长度限定，列出目标基因的在多少bp之间。 与模板基因比较，确定是否属于所需要的基因。例如选取一个rbcL基因的模板，当从ncbi下载数据或者用户上传的数据的时候，系统用这个模板来比较。 假基因的筛选（这一步我们平台都做了） 原文：All submitted sequences are first translated into amino acids and are compared against a Hidden Markov Model of the COI protein to verify that they actually derive from COI. Sequences that pass this check are then examined for stop codons(to detect possible pseudogenes) and are also compared against a small suite of possible contaminants (e.g. human). 我们能否给rbcl、matk、atpb分别选取模板来比对？ 关于BLAST抽提应用 我看到一篇文献Any query sequence is aligned very quickly to the global alignment through a Hidden Markov Model (HMM) profile of the COI protein (Eddy 1998), followed by a linear search of the reference library. Blast methods were tested, but users had difficulty interpreting results because scores are influenced by sequence length as well as by sequence similarity. Moreover, the best blast hit is of little value when no closely related taxa are in the reference database (Pertsemlidis Fondon 2001). Finally, benchmark studies (data not shown) demonstrated that linear searches supported by HMM alignments are faster than blast (which was primarily designed for local similarity searches (Altschul et al 1990). Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. Journal of Molecular Biology, 215, 403–410. Eddy SR (1998) Profile hidden Markov models. Bioinformatics,14, 755–763. Pertsemlidis A, Fondon JW (2001) Having a BLAST with Bioinformatics (and Avoidi