A linear-time algorithm for finding the longest segment which scores above a given threshold

TL;DR

This paper introduces a linear-time algorithm to identify the longest sequence segment with a sum or average above a specified threshold, applicable in DNA analysis and genome assembly preprocessing.

Contribution

It presents a novel linear-time algorithm for finding the longest segment with sum or average constraints, improving efficiency over previous methods.

Findings

Algorithm runs in linear time

Applicable to DNA sequence analysis

Useful for genome assembly preprocessing

Abstract

This paper describes a linear-time algorithm that finds the longest stretch in a sequence of real numbers (``scores'') in which the sum exceeds an input parameter. The algorithm also solves the problem of finding the longest interval in which the average of the scores is above a fixed threshold. The problem originates from molecular sequence analysis: for instance, the algorithm can be employed to identify long GC-rich regions in DNA sequences. The algorithm can also be used to trim low-quality ends of shotgun sequences in a preprocessing step of whole-genome assembly.