High-throughput immune repertoire analysis with IGoR

Abstract

High-throughput immune repertoire sequencing is promising to lead to new statistical diagnostic tools for medicine and biology. Successful implementations of these methods require a correct characterization, analysis, and interpretation of these data sets. We present IGoR (Inference and Generation Of Repertoires)—a comprehensive tool that takes B or T cell receptor sequence reads and quantitatively characterizes the statistics of receptor generation from both cDNA and gDNA. It probabilistically annotates sequences and its modular structure can be used to investigate models of increasing biological complexity for different organisms. For B cells, IGoR returns the hypermutation statistics, which we use to reveal co-localization of hypermutations along the sequence. We demonstrate that IGoR outperforms existing tools in accuracy and estimate the sample sizes needed for reliable repertoire characterization.

Predicting the spectrum of TCR repertoire sharing with a data‐driven model of recombination

“…Our prediction for sharing is mainly based on the generation model,29 which is sequence specific, attributing each sequence its own probability of generation. We have found that an overall selection factor is needed to predict sharing numbers correctly, but this simple and effective model is sequence independent.…”

“…This task is made difficult by the fact that, as a consequence of convergent recombination, the specific recombination event behind an observed sequence is not directly accessible. However, methods of statistical inference can be used to overcome this problem and learn accurate models of V(D)J recombination,26, 27, 29, 39 models which can in turn be used to predict sharing properties of sampled repertoires or of individual TCR sequences. These models have been shown to vary little between individuals, with small differences only in the germline gene usage and remarkable reproducibility in the insertion and deletion profiles 26.…”

“…To obtain model predictions for humans, we used a previously described model for TCRβ sequence generation inferred by the software package IGoR29 from repertoire data of a single individual 30. IGoR infers the probability distribution of V(D)J recombination events from sequence data (see details in Section 7.1).…”

“…To calculate the generation probability of TCRs, one needs to sum the occurrence probabilities of all the possible recombination events leading to a given nucleotide sequence26, 29 and, since we choose to follow CDR3 amino acid sequences, sum the probabilities of all nucleotide sequences leading to the amino acid sequence of interest. This is a computationally hard task that can be rendered tractable using a dynamic programming approach (see Section 7).…”

“…Before discussing the performance of this classifier, it is important to note that it is based on a model of recombination trained in a completely unsupervised way, ie, without using any information about the public status of the sequences. In fact, this training can be done with IGoR 29 from the repertoire of a single individual, without including any sharing information. Unlike previous approaches,23 we do not fit additional model features based on the catalog of sequences with their public or private status.…”