# Workshop on Data Structures in Bioinformatics - Program

## 1 Final program

Each talk lasts 30 min + 10 min questions. See titles and abstracts in the next section below.

• Monday 8th December
• Welcome 9:00
• Session 1: 9:30 - 12:30
1. B. Cazaux
2. S. Puglisi
3. N. Prezza
• Break
• Session 2: 14:30 - 17:45
1. G. Rizk
2. P. Melsted
3. 15h50-16h15 Coffee
4. G. Holley
5. Discussion
• Tuesday 9th December
• Session 1: 9:00 - 12:15
1. A. Héliou
2. D. Belazzougui
3. 10h20-10h50 Coffee
4. J. Köster
5. P. Bonizzoni
• Break
• Session 2: 14:00 - 17:00
1. T. Lecroq
2. G. Kucherov
3. 15h20 - 15h50 Coffee
4. S. Pissis
5. A. Schoenhuth
6. Discussion and conclusions

## 2 Titles and abstracts

• Simon Puglisi, Univ Helsinki, Finland
• Title: Scalable Indexing of Highly Repetitive Data
• Abstract: Advances in DNA sequencing mean databases of thousands of human genomes will soon be commonplace. This talk will introduce a simple technique for reducing the size of conventional indexes on such highly repetitive texts. Given upper bounds on pattern lengths and edit distances, we preprocess the text with LZ77 to obtain a filtered text, for which we store a conventional index. Later, given a query, we find all matches in the filtered text, then use their positions and the structure of the LZ77 parse to find all matches in the original text. Our experiments show this also significantly reduces query times.

• Pall Melsted, University of Iceland, Iceland, pmelsted@hi.is
• Title: Bloom filters for fun and profit
• Abstract: A Bloom filter is a simple, fast and memory efficient probabilistic data structure for storing a set. The Bloom filter has a one sided error, in the sense that it can give false positive answers to membership, but the error rate is controlled by the amount of memory used. We will review the use of Bloom filters in applications such as k-mer counting, assembly and error correction for HTS datasets. Additionally we will cover some variations of Bloom filters and introduce a new variant with better memory efficiency for storing k-mers.

• Guillaume Rizk, INRIA, France, guillaume.rizk@inria.fr
• Title: k-mer counting with minimizers
• Abstract: K-mer counting is a recurrent step in many NGS algorithms ( assembly, read error correction, …). Current approaches are either ram-intensive (Jellyfish), or disk-based with huge amount of input/output to disk (DSK, KMC). A new disk-based method based on minimizers introduced by Deorowicz et al., reduces the amount of I/O by an order of magnitude, providing both memory frugal and fast k-mer counting. The GATB library (Genome Assembly & Analysis Tool Box https://gatb.inria.fr) now includes an implementation of k-mer counting based on those ideas. This talk will detail the counting procedure, and comparative results.

• Nicola Prezza and Alberto Policriti, Univ. Udine, Italy
• Title: Hashing and indexing with succinct hash data structures
• Abstract: In this talk, we will begin by illustrating how we integrated hashing and indexing in building an aligner (the ERNE aligner) used to map short patterns (reads) on a long text (reference) with a limited amount of errors. Our hash-based strategy builds upon two families of hash functions. The first, named Hamming-aware, permits to considerably reduce search space by hashing the entire Hamming ball around the searched pattern to a smaller one. The second, named de Bruijn, allows to obtain a novel kind of succinct hash data structure whose space requirements are of n+o(n) bits instead of O(n log n). We will conclude showing how we obtained new space-time bounds for the best k-mismatch problem by finding a hash function with both the above described properties.

• Johannes Köster and Sven Rahmann, Univ. Essen, Germany
• Title: Massively parallel read mapping with the q-group index and PEANUT
• Abstract: We present the q-group index, a novel data structure for read mapping tailored towards graphics processing units (GPUs) with a small memory footprint and efficient parallel algorithms for querying and building. On top of the q-group index we introduce PEANUT, a highly parallel GPU-based read mapper.

• Bastien Cazaux, Thierry Lecroq and Eric Rivals
• Title: From indexing data structures to de Bruijn graphs
• Abstract: New technologies have tremendously increased sequencing throughput compared to traditional techniques, thereby complicating DNA assembly. Hence, assembly programs resort to de Bruijn graphs (dBG) of $$k$$-mers of short reads to compute a set of long contigs, each being a putative segment of the sequenced molecule. Other types of DNA sequence analysis, as well as preprocessing of the reads for assembly, use classical data structures to index all substrings of the reads. It is thus interesting to exhibit algorithms that directly build a dBG of order $$k$$ from a pre-existing index, and especially a contracted version of the dBG, where non branching paths are condensed into single nodes. Here, we formalise the relationship between suffix trees/arrays and dBGs, and exhibit linear time algorithms for constructing the full or contracted dBGs. Finally, we provide hints explaining why this bridge between indexes and dBGs enables to dynamically update the order $$k$$ of the graph.

• Gregory Kucherov, Univ. Marne La Vallée, France
• Title: Improved filters for the approximate suffix-prefix overlap problem
• Abstract: The approximate suffix-prefix overlap problem is to find all pairs of strings from a given set such that a prefix of one string is similar to a suffix of the other. This computation is a basic operation in genome assembly (overlap-layout-consensus approach) and in read correction (multiple sequence alignment approach). Valimaki et al. (Information and Computation, 2012) gave a solution to this problem based on suffix filters. In this work, we propose two improvements to the method of Valimaki et al. that reduce the running time of the computation.

This is a joint work with Dekel Tsur (Ben-Gurion University).

• Thierry Lecroq, Univ. Rouen, France
• Title: Fast on-line pattern matching algorithm for highly similar sequences
• Abstract: With the advent of NGS technologies there are more and more genomic sequences of individuals of the same species available. These sequences only differ by a very small amount. There is thus a strong need for efficient algorithms for performing fast pattern matching in such specific sets of sequences. We will present very efficient algorithms that solves the on-line exact pattern matching problem in a set of highly similar DNA sequences. The algorithm we propose extends the Morris-Pratt and Knuth-Morris-Pratt algorithms and variants of the Boyer-Moore exact string matching algorithm. Experimental results show that our new algorithms exhibit good performances in practice.

This is a joint work with N. Ben Nsira and M. Elloumi.

• Solon Pissis, Kings College, London
• Title: Accurate and efficient methods to improve multiple circular sequence alignment
• Abstract: Multiple sequence alignment is a core computational task in bioinformatics and has been extensively studied over the past decades. This computation requires an implicit assumption on the input data: the left- and right-most position for each sequence is relevant. However, this is not the case for circular structures; for instance, MtDNA. Efforts have been made to address this issue but it is far from being solved. We have very recently introduced a fast algorithm for approximate circular string matching (Barton et al., Algo Mol Biol, 2014). Here we first show how to extend this algorithm for approximate circular dictionary matching; and, then, apply this solution with agglomerative hierarchical clustering to find a sufficiently good rotation for each sequence. Furthermore, we propose an alternative method that is suitable for more divergent sequences. We implemented these methods in BEAR, a programme for improving multiple circular sequence alignment. Experimental results, using real and synthetic data, show the high accuracy and efficiency of these new methods in terms of the inferred likelihood-based phylogenies.

This is a joint work with Carl Barton, Costas S. Iliopoulos, and Fatima Vayani.

• Djamal Belazzougui, Univ Helsinki, Finland
• Title: BW4SA A sequence analysis library based on the Burrows-Wheeler transform
• Abstract: Most sequence analysis problems can be solved with the help of the suffix tree, probably the most important and well-known text indexing data structure. The main problem with the suffix tree is its excessive space usage. Given a text T length n over an alphabet of size sigma a suffix tree occupies O(n log n) bits of space while the text itself occupies only nlog sigma bits of space. The suffix tree can be used to the sequence analysis problems in O(n log sigma) time. The compressed suffix tree (Sadakane TOCS 2007) is a relatively new data structure which uses only $$O(n \log \sigma)$$ bits of space, and can simulate any algorithm on the suffix tree at the price of a slowdown by a factor $$O((\log n)^\epsilon)$$ for any constant $$\epsilon > 0$$. Recently, Beller, Berger, Gog, Ohlenusch and Schnattinger (SPIRE 2011,2012, JDA 2013) presented a general algorithm based on Burrows-Wheeler transform (BWT) that can be used to find maximal repeats and solve other sequence analysis problems in $$O(n \log sigma)$$ time and bits of space, thus achieving the speed of the suffix tree and the space-efficiency of the compressed suffix tree. The general idea of the new algorithm is to enumerate the nodes of the suffix tree by navigating the suffix link tree (SLT). In a recent paper, Belazzougui, Cunial, Karkkainen and Makinen (ESA 2013) have shown a different algorithm to traverse the SLT the $$O(n \log \sigma)$$ time, given the bidirectional BWT of a sequence and use it to solve many other sequence analysis problems. In this talk i will present BW4SA (Burrows-Wheeler for sequence analysis) a sequence analysis library that implements many of the algorithms presented in (ESA 2013).

This is a joint work in progress with Fabio Cunial. Implementation by Topi Paavilainen, Paula Lehtola, Max Sandberg, Lassi Vapaakallio, Jarno Alanko and Ahmed Sobih.

• Alice Heliou, Polytechnique, LIX, Paris
• Title: Linear-time Computation of Minimal Absent Words Using Suffix Array
• Abstract : An absent word of a word y of length n is a word that does not occur in y. It is a minimal absent word if all its proper factors occur in y. Minimal absent words have been computed in genomes of organisms from all domains of life; their computation also provides a fast alternative for measuring approximation in sequence comparison. In this talk we will present the first O(n)-time and O(n)-space algorithm for computing all minimal absent words based on the construction of suffix arrays. Experimental results, using real and synthetic data, show that our implementation outperforms the existing one.

This is a joint work with Carl Barton, Laurent Mouchard and Solon Pissis.

• Paola Bonizzoni, Università Degli Studi di Milano-Bicocca, Italy
• Title: An external memory BWT-based assembler for Next Generation data
• Abstract: The large amount of short read data that are needed for assembling in some fields (e.g. metagenomic) strongly motivates the investigation of disk-based approaches to indexing reads. Positive results in this direction can stimulate the investigation of efficient disk-based algorithms for de novo assembly of reads. We developed a disk based-algorithm for building string graphs and we have integrated our implementation into SGA (the most well-known assembler based string graphs). Besides the integration with an existing assembler, the main advantages of our algorithm are (1) a main memory requirement independent of the size of the data set, and (2) a running time that is competitive with the (memory-based) one of SGA. We have performed an experimental analysis to determine the current limitations of an external memory assembler, starting form the state of the art of building an FM-index and LCP, Suffix Array in external memory.

This is a joint work with Marco Previtali, Università Degli Studi di Milano-Bicocca, Italy.

• Guillaume Holley, Univ. Bielefeld, Germany
• Title: Bloom Filter Trie - a data structure for pan-genome storage
• Abstract: High Throughput Sequencing technologies have become fast and cheap in the past years. As a result, large-scale projects started to sequence tens to several thousands of genomes per species. The concept of pan-genome has therefore emerged. It is composed of two parts: The core genome, which represents the pool of all the genes shared by all the strains of the species, and the dispensable genome, which represents genes not shared by all strains. Graphs exhibit good properties for representing a pan-genome, as they avoid storing redundant information and can represent genomic variants. Unfortunately, existing tools using this representation are slow or memory consuming. We present here a new data structure for storing pan-genomes: The Bloom Filter Trie. This data structure allows to store a pan-genome as an annotated de-Bruijn graph in a memory efficient way.

This is a joint work with Roland Wittler and Jens Stoye from International Research Training Group 1906/1 and Genome Informatics, Faculty of Technology, Bielefeld University, Germany.

• Alexander Schoenhuth, CWI, Amsterdam, Holland
• Title: Genome sequence assembly via maximal clique enumeration
• Abstract: Genomes of unknown ploidy and high diversity, such as those of many pathogens, can harbor haplotypes of (very) low frequency. While one can address this by deep sequencing, the elevated error rates that apply for next-generation sequencing can interfere with the low frequency of the variants inherent to the low-frequency haplotypes. New ideas for error correction and haplotype-aware assembly are required when assembling such highly diverse, polyploid genomes. We present a new method, HaploClique, which is based on enumerating maximal cliques in overlap-like graphs as an underlying principle. In these graphs, nodes reflect probability distribution over sequences, which at the beginning are given by the reads, and edges reflect that two such distributions are likely to have arisen from the same underlying base sequence, i.e. from (locally) the same haplotype. Maximal cliques then reflect maximal ensembles of reads likely to stem from the same haplotype. One then iterates by transforming max-cliques into nodes and drawing the corresponding edges among those new nodes, which results in growing, error-corrected haplotype contigs. As proof of principle, we show that we can reliably assemble HIV genomes – we outperform all prior, state-of-the art methods that aim at global assembly of highly diverse, polyploid genomes.

This is joint work with Tobias Marschall, MPI Informatics, Saarbruecken, Armin Toepfer and Niko Beerenwinkel from ETHZ, Basel.