VAliBS: a visual aligner for bisulfite sequences

Min Li, Ping Huang, Xiaodong Yan, Jianxin Wang, Yi Pan, Fang-Xiang WuBMC Bioinformatics 2017 · OpenReview

TL;DR

This paper introduces VAliBS, a novel visual tool for aligning bisulfite-treated DNA sequencing reads to a reference genome. VAliBS addresses the challenges of mapping these reads, which undergo chemical modifications, by employing a "three-letter" strategy and integrating two popular aligners (Bowtie2 and BWA), offering improved accuracy and a user-friendly visualization of methylation patterns.

What Problem Does This Paper Address?

DNA methylation is a crucial epigenetic modification linked to gene expression and various diseases, making its study a hot topic in medical science. However, a common method for detecting methylation, bisulfite treatment, chemically converts unmethylated cytosines (C) to uracils (U), which are then read as thymines (T) during sequencing. This chemical transformation breaks the standard rules of DNA sequence alignment, where C should only match C and T should only match T. Traditional aligners, designed for unmodified DNA, struggle with these bisulfite-converted reads, leading to inaccurate mapping and difficulty in pinpointing methylation sites. The paper aims to solve this by developing a specialized aligner that can accurately map these modified sequences.

Methods

VAliBS employs a three-stage pipeline:

  1. Pre-processing: This stage is crucial for handling the C-to-T conversion. It involves creating two versions of both the reference genome and the sequencing reads:
  2. One set converts all 'C's to 'T's (read_c2t, ref_c2t).
  3. Another set converts all 'G's to 'A's (read_g2a, ref_g2a) to account for the reverse complement strand's conversion. This effectively masks the difference between original 'C's and 'T's, allowing standard aligners to work.

  4. Mapping: VAliBS integrates two established mapping tools, Bowtie2 and BWA. Users can choose to use either or both, providing flexibility. The pre-processed reads are then mapped to the pre-processed reference genomes using these tools.

  5. Post-processing: This stage filters out false mappings that might arise from the pre-processing step. It also considers Single Nucleotide Polymorphism (SNP) data to avoid incorrectly flagging known variations as mapping errors. Finally, it merges results from Bowtie2 and BWA (if both were used) and removes redundant mappings. A key feature is the visualization of alignment results, highlighting methylated sites (marked in red) and unmethylated sites (marked in green) using an X-Window interface.

Datasets

The paper evaluates VAliBS using:

  • Simulated datasets: These were generated using the ART simulator, with reads derived from the human hg19 chr22 reference. Both error-free and error-containing simulated reads (75 bp and 100 bp) were created, with random C/G conversions based on known human DNA methylation levels.
  • Real datasets: Single-end reads from SRR299053 (mouse) and paired-end reads from SRR306438 (human) were used.

The paper does not introduce new benchmark datasets but compares its performance against existing tools on these standard simulation and real-world datasets.

Key Insights and Ideas

  • The "three-letter" strategy is effective: By creating modified versions of both reads and the reference, VAliBS cleverly adapts standard mapping tools to the unique challenges of bisulfite sequencing.
  • Integration of multiple aligners offers robustness: Combining Bowtie2 and BWA allows VAliBS to leverage the strengths of each, potentially improving recall and precision, especially when dealing with noisy data.
  • Visualization is a game-changer: The intuitive graphical interface that highlights methylation status directly on the alignment is a significant usability improvement over command-line tools, making it easier for researchers to interpret results.
  • Post-processing is critical for accuracy: Filtering out false positives introduced by the base conversion process is essential for reliable methylation calling.

Key Results

  • Improved Accuracy: VAliBS demonstrated improved accuracy in bisulfite mapping compared to state-of-the-art tools like BisMark, BSMAP, and BS-Seeker2, particularly on paired-end data.
  • Higher Mappability on Real Data: On real datasets, VAliBS consistently achieved higher mappability (percentage of reads that can be mapped) than Bismark, BS-Seeker2, and BSMAP for both single-end and paired-end reads.
  • User-Friendly Interface: The visual interface allows for easy operation and clear visualization of alignment details and methylation positions.
  • Flexibility: VAliBS supports various input formats and can integrate with both Bowtie2 and BWA, offering users flexibility in parameter choices.

Limitations and Future Work

  • Limitations: The paper notes that for real data, it's difficult to definitively confirm the correctness of unique mappings, so comparisons are primarily based on mappability. The tool's performance on RRBS data was not specifically tested as it didn't include special treatments for it.
  • Future Work: The authors suggest that the efficiency of VAliBS could be further improved by leveraging parallel processing techniques like MapReduce, given its modular design.

Summary updated

Abstract

Methylation is a common modification of DNA. It has been a very important and hot topic to study the correlation between methylation and diseases in medical science. Because of the special process with bisulfite treatment, traditional mapping tools do not work well with such methylation experimental reads. Traditional aligners are not designed for mapping bisulfite-treated reads, where the un-methylated ‘C’s are converted to ‘T’s. In this paper, we develop a reliable and visual tool, named VAliBS, for mapping bisulfate sequences to a genome reference. VAliBS works well even on large scale data or high noise data. By comparing with other state-of-the-art tools (BisMark, BSMAP, BS-Seeker2), VAliBS can improve the accuracy of bisulfite mapping. Moreover, VAliBS is a visual tool which makes its operations more easily and the alignment results are shown with colored marks which makes it easier to be read. VAliBS provides fast and accurate mapping of bisulfite-converted reads, and a friendly window system to visualize the detail of mapping of each read. VAliBS works well on both simulated data and real data. It can be useful in DNA methylation research. VALiBS implements an X-Window user interface where the methylation positions are visual and the operations are friendly.
BibTeX
@article{li2017valibs,
  author = {Min Li and Ping Huang and Xiaodong Yan and Jianxin Wang and Yi Pan and Fang-Xiang Wu},
  title = {{VAliBS: a visual aligner for bisulfite sequences}},
  journal = {BMC Bioinformatics},
  year = {2017},
  url = {https://openreview.net/forum?id=0XXA45WAIm},
}
Figure