Optimising Nanopore Sequencing for Forensic STR Analysis

Abstract

Nanopore sequencing for forensic analyses provides several advantages including portability, real-time analysis, and reduced costs, but this has yet to be optimised for low-input samples commonly encountered in forensic casework. Part one of this study aimed to develop a forensic STR analysis workflow able to provide correct genotyping results using nanopore sequencing. Optimisations to the workflow included enrichment using whole genome amplification (WGA) to improve DNA yield of low-input samples. Different library preparation methods, flow cells, and bioinformatic tools were also investigated. Once a STR analysis workflow was developed, part two of this study evaluated the workflows further using more challenging samples, including low-input, forensic-like, and mixture samples. Part one of this study demonstrated correct genotyping of 12 out of 14 STR loci using no enrichment techniques, with FGA and CD4 being identified as loci prone to incorrect genotyping due to homopolymer sequences in the repeat structures. All three whole genome amplification methods evaluated provided sufficient increases in DNA yield, with the REPLI-g® Single Cell Kit generating accurate profiles comparable to no enrichment samples. R10.3 flow cells could not be fully assessed for their improvement in sequencing homopolymers potentially due to incompatibility of the sequencing chemistry with the library preparation method chosen. Two MPS tools, Fragsifier and STRait Razor, were evaluated for STR sequence extraction using nanopore data, with Fragsifier generating profiles with improved read counts with higher accuracy. Part two of this study showed that the no enrichment workflow can generate full STR profiles from as low as 100 pg of starting material, with partial profiles using 50 pg. The use of WGA on lowinput samples introduced additional challenges due to the higher levels of stochastic effects observed such as amplification bias and heterozygote imbalance. These effects were exacerbated in mixture samples using either the no enrichment or enrichment with WGA workflow, and the observed mixture ratio could not be used to accurately predict the expected mixture ratio in any sample analysed. This study highlights the limitations of this sequencing platform and further improvement is required before nanopore sequencing can be considered for a forensic purpose.