Comparing the performance of sequence alignment software : Bandage, SPAligner, GraphAligner

Abstract

In this thesis, we investigate the performance of available methods and tools for sequence alignment in assembly and de Bruijn graphs. Sequence alignment tools are employed to detect antimicrobial (AMR) gene sequences within the assembly graphs. Utilizing precise and efficient tools for identifying these genes enables us to locate their neighboring genes and evidence of horizontal gene transfer (HGT) more accurately. To this end, we have considered three sequence alignment tools namely Bandage, SPAligner and GraphAligner. The tools have similar input and output types. The outputs are analyzed qualitatively and quantitatively using Panda, Numpy and GFA libraries in Python. The paths returned by each pair of tools for each query are compared to measure the similarity between them. Furthermore, the output sequences from each software are compared to the target sequence using a modified version of edit distance. It was seen that Bandage was the most efficient and precise tool, followed by GraphAligner and then SPAligner for the datasets tested.