文章： UCHIME improves sensitivity and speed of chimera detection 嵌合体产生的原因，主要是PCR的过程： The most common mechanism is incomplete template extension, when a partially extended sequence from one sequence reanneals to another parent in the next cycle of PCR. This problem is particularly acute in population studies that sequencea single region, such as the bacterial 16S ribosomal RNA gene(16S) or the fungal Internal Transcribed Spacer (ITS) region. 嵌合体的种类一般有： While chimeras with two segments ( bimeras ) are most common, chimeras with 2 segments ( multimeras ) may form at comparable rates and account for a significant fraction of the unique sequencesin an amplified sample. 嵌合体检测的工具有： CHIMERA_CHECK、 Pintail、 Mallard、 Bellerophon、 ChimeraSlayer. 各自检测的原理如下， Pintail and Mallard are 16S-specific programs that use a reference database of trusted chimera-free reference sequences. The query sequence is aligned to all( Pintail ) or all pairs ( Mallard ) of reference sequences. Evolutionary distance is computed in a sliding window across the query sequenceand variations in distance are compared with the known ratevariability in the 16S gene, with larger variations indicating achimera. ChimeraChecker is an ITS-specific method using BLAST to search a reference database for taxonomic anomalies. ChimeraSlayer searches a multiple alignment of chimera-free reference sequences and constructs three-way alignments with candidate parents. Perseus is designed todetect chimeras in 454 pyrosequencing reads that have been filtered by the AmpliconNoise algorithm . Assuming that a chimera has undergone fewer rounds of amplification than its parents, the query is compared with all pairs of sequences having higher abundance. 方法主要分两类：基于reference database的检测，基于 de novo 策略的检测。 UCHIME的检测包含两种模式， UCHIME can use a trusted reference database of non-chimeric sequences (like ChimeraSlayer) and also offers a de novo mode (like Perseus). UCHIME基于reference database的检测策略，主要原理如下图： 此算法主要分为以下几步， 1） The query sequence is divided into four non-overlapping segments ( chunks ), each of which is used to search a reference database, which is assumed to be chimera free. （将query sequence拆分成不存在overlap的chunks，然后比对数据库） 2） The best matches to each chunk are noted, and the two best candidate parents are identified from matches to all chunks. （选取每个chunk在数据库中最好的match，最终选择两条最好的parents序列） 3） A three-way multiple alignmentof the query to these two candidates is constructed. 4） If a pair of segments extracted from these two candidates has identity ≥ 0 . 8% closer to the query sequence than either candidate alone, a score is computed from the alignment and a chimera is reported if the score exceeds a predetermined threshold. 其中 ，UCHIME selects the best two candidates according to the following procedure，对于选择最好的parents，操作如下： A pair-wise alignment is computed between the query Q and each candidate parent P. The identity between P and Q is smoothed over a window(default size 32). For each position in Q, the highest value for the smoothed identity among the parents is recorded. The best candidate is then identifiedas the one with most positions having highest smoothed identity. 对于three-way的alignment，比对会出现local、local-X和global-X，三种情况如下： 这三种情况的操作如下：(i) searching for global-X alignments, as fewer global-X alignments usuallyexist compared with local or local-X; (ii) examining only two candidateparents; and (iii) discarding models having distance to the closest parent(divergence) 0.8%.最终会形成如下比对结果： 根据此比对结果进行score值计算，分值计算公式如下： h是界限值， g可以表示L(left)或者R(right)，Let Yg, Ng and Ag be the total numberof yes, no and abstain votes in segment g of the model.最终依据此分值去判段序列是不是嵌合体序列。 UCHIME基于de novo的检测策略，主要原理如下： sequences are considered in the order of decreasing abundance,and candidate parents must have abundance at least 2 × that of the query sequence, assuming chimeras are less abundant than their parents because they undergo fewer rounds of amplification. 主要操作是 Candidate parents are required to have abundance at least λ times that of the query sequence, on the assumption that a chimera has undergone fewer rounds of amplification and will therefore be less abundant than its parents. The parameter λ is called the abundance skew , and by default λ = 2,assuming at least one more round of amplification for the parents. 主要是通过量上的变化来进行检测。 UCHIME的最终检测效果如下： UCHIME检测的敏感性相较其他数据较好，速度更快！ 参考文章： http://www.ncbi.nlm.nih.gov/pubmed/21700674
在16s rRNA的分析中，在数据处理过程中重要的一步操作是鉴定和去除嵌合体。 嵌合体（chimera）产生的原因主要是PCR过程中产生的错误： During this PCR amplification，chimeras might be created due to incomplete extension. 在扩增过程中，chimera的比例可能会达到70% ： Likewise, the percentage of chimeric se-quences in the unique amplicon pool of PCR-amplified samplesmight reach values higher than 70%。（ 实验过程的优化中，考虑减少嵌合体的产生 ）。 嵌合体的处理策略，主要可分 为 reference-based和 de novo 两种。 reference-based的原理是 Reference-based methods basically screen the sequences poten-tially containing chimeras against a curated reference databasewith chimera-free sequences. 工具有 Pintai l 、 Bellerophon 。在这基础上 ChimeraSlayer 实现了较大的改动和性能优化， ChimeraSlayer的基本原理是 which uses 30% of each end as a seed forsearching a reference data set, finding the closest parent (if any)，performing alignments, and scoring to the candidate parents. 它的缺点在于 it was not able to detect chimeras with a smallchimeric range. 在 ChimeraSlayer的基础 上， reference-based UCHIME 表现性能更好，In reference-based UCHIME, query sequences are divided into four nonoverlapping segments andsearched against a reference database. 有研究报道，ChimeraSlayer and reference-based UCHIME在长reads中具备短chimeric的时候效果不如DECIPHER，were reported to have a lower accuracy than that of DECIPHER in cases where the algorithms were challenged with a data set con-taining chimeric sequences with a short chimeric range and longsequence lengths. DECIPHER的原理是The DECIPHER algorithm is a search-based algorithm that splits the query sequence into different fragmentsand analyzes whether those fragments are uncommon in the ref-erence phylogenetic group where the query sequence is classified. If a significant amount of fragments is assigned to a phylogeneticgroup different from the complete query sequence, the sequence isclassified as chimeric. 实际上，chimera检测工具的性能评估很难做到公平统一，各工具有各自的适应范围。 De novo 策略的原理是 De novo methodologies are generally based on the fact thatparents of any chimeric sequence have gone through at least onemore PCR cycle than chimeric sequences. 工具有 Perseu s 、 de novo UCHIME 、 de novo ChimeraSlayer ，这些工具目前都已整合到了mothur中。 近来 the UPARSE pipeline was released, combining in one step chimera detection with clus-tering of sequencing reads into operational taxonomic units. reference-based和 de novo 两种策略各自具有优缺点：1）reference-based的优势在于 In situations dealing with well-studied environments, the reference-based approaches werefound to be very effective in distinguishing between chimeras andchimera-free (parent) sequences . 2）reference-based的劣势在于 efficiency is assumed to belower when dealing with less well-known environments.而这正好是 de novo 方法的优势所在。3） de novo 方法的劣势在于 most of the de novo approaches depend on redundancy differences between chimeras and parents, assuming that the number of parentsequences has to be at least one time more redundant than theircorresponding chimeric sequences. This requires data abundances to have been reported with high accuracy. （这个就涉及到多少数据量是能保证效果的） 15年CATCh（Combining Algorithms to Track Chimeras）出现，其原理在于利用其他chimera的检测工具作为input，利用有监督的学习方法去进行分类模型构建，利用测试集验证分类模型的准确度，最终确定分类模型来鉴定嵌合体： which is able to discriminate betweenchimeric and nonchimeric sequences based on a specific set ofinput data (called features in the context of machine learning).在此工具中，输入数据不是测序reads，而是不同工具鉴定chimera的结果。 With this tool, we use as input data not the sequence read charac-teristics but rather the results (e.g., scores) of different individual chimera detection tools mentioned above and integrate them intoone prediction. All different tools are run separately, and theiroutput values are combined and processed by the classifier in or-der to give a prediction of whether a read is a chimera or not. 此工具在处理时，主要分为3个步骤：（1） the necessary input features (i.e., output values of the different chimeradetection tools) are identified. （2） the classifier istrained via a supervised learning approach. In this step, the classifier learns to make a correct prediction based on example input data; in our case, training data consist of the output features of a set of sequences reads obtained from different chimera detection tools, together with their correct classification (i.e., whether thisread is a chimeric sequence or not). （3） In the third step, the trained classifier can be used to predict chimeric sequences in new, previously unseen data (i.e., data that did not belong to the training data). By feeding the outputs of the different individual chimeradetection tools into the classifier, CATCh is able to classify them into chimeric and chimera-free subsets. As two different types ofchimera detection tools exist, either reference based or de novo , wealso developed two different versions of CATCh. In order to illus-trate its performance, CATCh (reference based as well as de novo )was benchmarked against other chimera detection tools using var-ious publicly available benchmark data. （ 利用其他工具的检测结果作为输入数据，不同工具的结果出现不一致的情况时，对模型结果是否存在影响 ） 参考文献： Mysara M, Saeys Y, Leys N et al. CATCh, an ensemble classifier for chimera detection in 16S rRNA sequencing studies. 2015, 81(5):1573-84. doi: 10.1128/AEM.02896-14
标签: 嵌合体
