Sketching and Sequence Alignment: A Rate-Distortion Perspective

TL;DR

This paper introduces a rate-distortion framework for DNA read sketching, proposing a new locational hashing algorithm that reduces sketch size requirements and enhances computational efficiency in pairwise alignment tasks.

Contribution

It develops a novel rate-distortion approach and a new locational hashing algorithm that significantly reduces sketch size compared to standard methods.

Findings

Locational hashing requires fewer bits than min-hash for the same distortion.

The proposed method achieves a logarithmic squared dependence on inverse distortion.

Significant computational savings are possible in DNA sequence alignment.

Abstract

Pairwise alignment of DNA sequencing data is a ubiquitous task in bioinformatics and typically represents a heavy computational burden. A standard approach to speed up this task is to compute "sketches" of the DNA reads (typically via hashing-based techniques) that allow the efficient computation of pairwise alignment scores. We propose a rate-distortion framework to study the problem of computing sketches that achieve the optimal tradeoff between sketch size and alignment estimation distortion. We consider the simple setting of i.i.d. error-free sources of length $n$ and introduce a new sketching algorithm called "locational hashing." While standard approaches in the literature based on min-hashes require $B = (1/D) \cdot O\left( \log n \right)$ bits to achieve a distortion $D$, our proposed approach only requires $B = \log^2(1/D) \cdot O(1)$ bits. This can lead to significant computational savings in pairwise alignment estimation.