Efficient COI Barcoding Using High Throughput Single-End 400 bp Sequencing

Authors

• Chentao Yang, BGI-Shenzhen
• Yuxuan Zheng, Capital Normal University
• Shangjin Tan, BGI-Shenzhen
• Guanliang Meng, BGI-Shenzhen
• Wei Rao, BGI-Shenzhen
• Caiqing Yang, Capital Normal University
• David G. Bourne, James Cook University
• Paul A. O'Brien, James Cook University
• Junqiang Xu, BGI-Shenzhen
• Sha Liao, BGI-Shenzhen
• Ao Chen, BGI-Shenzhen
• Xiaowei Chen, BGI-Shenzhen
• Xinrui Jia, Capital Normal University
• Ai-bing Zhang, Capital Normal University
• Shanlin Liu, BGI-ShenZhen

Document Type

Article

Date of this Version

2020

Citation

BMC Genomics (2020) 21: 862

doi: 10.1186/s12864-020-07255-w

Comments

Open access material

License: CC BY 4.0 International

Abstract

Background

Over the last decade, the rapid development of high-throughput sequencing platforms has accelerated species description and assisted morphological classification through DNA barcoding. However, the current highthroughput DNA barcoding methods cannot obtain full-length barcode sequences due to read length limitations (for example, a maximum read length of 300 bp for the Illumina’s MiSeq system), or are hindered by a relatively high cost or low sequencing output (e.g. a maximum number of eight million reads per cell for the PacBio’s SEQUEL II system).

Results

Pooled cytochrome c oxidase subunit I (COI) barcodes from individual specimens were sequenced on the MGISEQ-2000 platform using the single-end 400 bp (SE400) module. We present a bioinformatic pipeline, HIFI-SE, that takes reads generated from the 5′ and 3′ ends of the COI barcode region and assembles them into full-length barcodes. HIFI-SE is written in Python and includes four function modules of filter, assign, assembly, and taxonomy. We applied the HIFI-SE to a set of 845 samples (30 marine invertebrates, 815 insects) and delivered a total of 747 fully assembled COI barcodes as well as 70 Wolbachia and fungi symbionts. Compared to their corresponding Sanger sequences (72 sequences available), nearly all samples (71/72) were correctly and accurately assembled, including 46 samples that had a similarity score of 100% and 25 of ca. 99%.

Conclusions

The HIFI-SE pipeline represents an efficient way to produce standard full-length barcodes, while the reasonable cost and high sensitivity of our method can contribute considerably more DNA barcodes under the same budget. Our method thereby advances DNA-based species identification from diverse ecosystems and increases the number of relevant applications.