Levenshtein error-correcting barcodes for multiplexed DNA sequencing

Abstract

BackgroundHigh-throughput sequencing technologies are improving in quality, capacity and costs, providing versatile applications in DNA and RNA research. For small genomes or fraction of larger genomes, DNA samples can be mixed and loaded together on the same sequencing track. This so-called multiplexing approach relies on a specific DNA tag or barcode that is attached to the sequencing or amplification primer and hence appears at the beginning of the sequence in every read. After sequencing, each sample read is identified on the basis of the respective barcode sequence.Alterations of DNA barcodes during synthesis, primer ligation, DNA amplification, or sequencing may lead to incorrect sample identification unless the error is revealed and corrected. This can be accomplished by implementing error correcting algorithms and codes. This barcoding strategy increases the total number of correctly identified samples, thus improving overall sequencing efficiency. Two popular sets of error-correcting codes are Hamming codes and Levenshtein codes.ResultLevenshtein codes operate only on words of known length. Since a DNA sequence with an embedded barcode is essentially one continuous long word, application of the classical Levenshtein algorithm is problematic. In this paper we demonstrate the decreased error correction capability of Levenshtein codes in a DNA context and suggest an adaptation of Levenshtein codes that is proven of efficiently correcting nucleotide errors in DNA sequences. In our adaption we take the DNA context into account and redefine the word length whenever an insertion or deletion is revealed. In simulations we show the superior error correction capability of the new method compared to traditional Levenshtein and Hamming based codes in the presence of multiple errors.ConclusionWe present an adaptation of Levenshtein codes to DNA contexts capable of correction of a pre-defined number of insertion, deletion, and substitution mutations. Our improved method is additionally capable of recovering the new length of the corrupted codeword and of correcting on average more random mutations than traditional Levenshtein or Hamming codes.As part of this work we prepared software for the flexible generation of DNA codes based on our new approach. To adapt codes to specific experimental conditions, the user can customize sequence filtering, the number of correctable mutations and barcode length for highest performance.

Enhancing the detection of barcoded reads in high throughput DNA sequencing data by controlling the false discovery rate

“…This property makes it suitable for use in DNA context (for full definitions and descriptions of this distance and algorithms described in this subsection see our previous work, [7]). The prefix of a sequence can be an empty sequence, the sequence itself, or any subsequence starting from position 1 up to any end position.…”

From next-generation resequencing reads to a high-quality variant data set

“…However, in the case of multiple samples having been sequenced simultaneously, which are distinguishable from one another by short, unique barcodes, these reads can cause more severe issues as it becomes necessary to demultiplex the read data in the presence of potential sequencing errors in these barcodes. To circumvent this problem, several methods have been developed for designing barcodes with an error-correction capability that aid correct sample identification in the presence of sequence alterations introduced during synthesis, primer ligation, amplification or sequencing (Buschmann and Bystrykh, 2013). Popular error-correcting techniques include methods based on false discovery rate statistics (see, for example, Buschmann and Bystrykh, 2013) as well as adaptations of both Hamming codes (see, for example, Hamady et al, 2008;Bystrykh, 2012) and Levenshtein codes (for example, implemented in the software Sequence-Levenshtein (Buschmann and Bystrykh, 2013) and TagGD (Costea et al, 2013)).…”

“…To circumvent this problem, several methods have been developed for designing barcodes with an error-correction capability that aid correct sample identification in the presence of sequence alterations introduced during synthesis, primer ligation, amplification or sequencing (Buschmann and Bystrykh, 2013). Popular error-correcting techniques include methods based on false discovery rate statistics (see, for example, Buschmann and Bystrykh, 2013) as well as adaptations of both Hamming codes (see, for example, Hamady et al, 2008;Bystrykh, 2012) and Levenshtein codes (for example, implemented in the software Sequence-Levenshtein (Buschmann and Bystrykh, 2013) and TagGD (Costea et al, 2013)). Algorithms based on the latter are capable of correcting not only substitution errors but also insertion and deletion (indel) errors that is particularly important for sequencing technologies where indels are the main source of error (that is, 454 and Ion Torrent).…”

DropSynth 2.0: high-fidelity multiplexed gene synthesis in emulsions

“…In order to generate distinct 12mer barcode sequences, we took 2,000 20mer primer sequences derived from Eroshenko et al 18 , removed all sequences containing NdeI, KpnI, BtsI-v2, BspQI, EcoRI, XhoI, SpeI, and NotI, and generated all possible 12mer subset sequences. We next screened for self-dimers, GC content between 45% and 55% and a melting temperature between 40°C and 42°C, We further filtered sequences to have a minimum modified Levenshtein distance of 3 between selected barcodes 27 . We then selected the first 384 sequences to be used in oligo designs, with complementary sequences being used to generate the beads.…”