However, codon frame and other effects have been observed to violate these parametric assumptions for such coding sequences, suggesting that a nonparametric approach to modeling the recombination process could be useful. Typical observed N-region lengths at the VD and DJ boundaries for two D and two J alleles.(TIFF) pcbi.1004409.s003.tiff (727K) GUID:?4EA8B591-C361-474A-9584-B7581FCF7815 S4 Fig: Mean and variance of inferred parameters. The across-subset mean and variance of inferred parameter values for each human in the Vollmers data set across 10 disjoint subsets of the data. See the caption to Fig 5 and the corresponding text for more details.(TIFF) pcbi.1004409.s004.tiff (468K) GUID:?AEF460BF-5463-438F-B01C-07147F2DAF6D S5 Fig: Deletions and N-region lengths for data and simulation for three different humans. (TIFF) pcbi.1004409.s005.tiff (468K) GUID:?A990634F-AA01-4E9C-82DC-F77256483F47 S6 Fig: Mutation frequencies for data and simulation for three different humans over the full reads, and for the V, D, and J segments individually. (TIFF) pcbi.1004409.s006.tiff (572K) GUID:?838B0AA5-9990-4A1B-91E2-9576EF5A9437 S7 Fig: Per-position mutation frequencies for data and simulation for typical alleles in the V, D, and J segments. (TIFF) pcbi.1004409.s007.tiff (1.0M) CycLuc1 GUID:?6541ADB2-7CA1-440A-903B-821B7E36BE33 Data Availability StatementThe data are available at the following dryad link: http://dx.doi.org/10.5061/dryad.149m8 Abstract VDJ rearrangement and somatic hypermutation work together to produce antibody-coding B cell CycLuc1 receptor (BCR) sequences for a remarkable diversity of antigens. It is now possible to sequence these BCRs in high throughput; analysis of these sequences is bringing new insight into how antibodies develop, in particular for broadly-neutralizing antibodies against HIV and influenza. A fundamental step in such sequence analysis is to annotate each base as coming from a specific one of the V, D, or J genes, or from an N-addition (a.k.a. non-templated insertion). Previous work has used simple parametric distributions to model transitions from state to state in a hidden Markov model (HMM) of VDJ recombination, and assumed that mutations occur via the same process across sites. However, codon frame and other effects have been observed to violate these parametric assumptions for such coding sequences, suggesting that a nonparametric approach to modeling the recombination process could be useful. In our paper, we find that CycLuc1 indeed large modern data sets suggest a model using parameter-rich per-allele categorical distributions for HMM transition probabilities and per-allele-per-position mutation probabilities, and that using such a model for inference leads to significantly improved results. We Ornipressin Acetate present an accurate and efficient BCR sequence annotation software package using a novel HMM factorization strategy. This package, called (https://github.com/psathyrella/partis/), is built on a new general-purpose HMM compiler that can perform efficient inference given a simple text description of an HMM. Author Summary The binding properties of antibodies are determined by the sequences of their corresponding B cell receptors (BCRs). These BCR sequences CycLuc1 are created in draft form by VDJ recombination, which randomly selects and deletes from the ends of V, D, and J genes, then joins them together with additional random nucleotides. If they pass initial screening and bind an antigen, these sequences then undergo an evolutionary process of mutation and selection, revising the BCR to improve binding to its cognate antigen. It has recently become possible to determine the BCR sequences resulting from this process in high throughput. Although these sequences implicitly contain a wealth of information about both antigen exposure and the process by which humans learn to resist pathogens, this information can only be extracted using computer algorithms. In this paper, we employ a computational and statistical approach to learn about the VDJ recombination process. Using a large data set, we find consistent and detailed patterns in the parameters, such as amount of V gene exonuclease removal, for this process. We can then use this parameter-rich model to perform more accurate per-sequence attribution of each nucleotide to either a V, D, or J gene, or an N-addition (a.k.a. non-templated insertion). Methods paper. [20] and the online annotation tool on the [21] website. Another approach has been to search sequences for motifs characteristic of various parts of the locus and search databases for the resulting segments [22]. However, BCR sequence formation is quite complex (reviewed in [11]) and this complexity invites a modeling-based approach, specifically in the framework of hidden Markov models (HMMs). HMMs for sequence analysis consist of a directed graph on hidden state nodes with defined start and end states, with each node potentially emitting a nucleotide base or amino acid residue [23, 24]. In the BCR case, the hidden states represent either (gene, nucleotide position) pairs or N-region nucleotides, and the emission probabilities incorporate the probability of somatic hypermutation at that base. The HMM approach to BCR annotation has been elegantly implemented first in [25], then [26], and then [27]. The transition probabilities for these previous HMM methods.