AID/APOBEC family represents a unique group of enzymes that function as DNA and RNA mutators. They can edit nucleotide sequence of DNA and/or RNA molecules thanks their ability to deaminate cytidine or its modified derivatives. In recent years, the detection of RNA/DNA editing events have been greatly improved, thanks the availability of high-throughput sequencing technologies. However, still bioinformatics analysis of next-generation sequencing (NGS) data sets remain a challenging task. Several pipelines are available for the detection of RNA editing events and experimentally induced mutations in genomic DNA. Accurate bioinformatics tools are essential to ensure the correct identification of edited sites. Below we present some of the most useful tools, given that in vitro activity of APOBECs causes slight nucleotide changes in RNA, and more global changes in DNA

In recent years, a number of software packages dedicated to transcriptome-wide identification of editing events have been developed. Most of them involve statistical models, filtering approaches, and integration of additional genomic information to improve the accuracy of editing detection and quantification. Some of these methods utilize RNA and matched DNA sequencing data from the same sample in order to reduce the identification of single-nucleotide polymorphisms and somatic mutations as editing events.

• REDItools consist of a set of scripts written in the Python language, available under the MIT license at https://github.com/BioinfoUNIBA/REDItools. REDItools are not organism oriented and work with RNA-Seq and DNA-Seq data from any sequencing platform. REDItools include three main scripts: REDItoolDnaRNA.py detects RNA editing candidates by comparing pre-aligned RNA-Seq and DNA-Seq reads; REDItoolKnown.py allows to explore the RNA editing potential of RNA-Seq experiments by looking at known events only; REDItoolDenovo.py performs the de novo detection of RNA editing candidates using RNA-Seq data alone. REDItools require inputs in BAM format, and results are provided in tab-formatted tables facilitating downstream analyses. It is highly flexible, including a variety of filters and quality checks, and may provide reliable sets of editing candidate sites.
• RES-scanner is a software package written in the Perl programming language, available under the GPL v3 license at https://github.com/ZhangLabSZ/RES-Scanner. RES-scanner provides a complete pipeline from raw sequencing reads to final editing sites and addresses read mapping, homozygous genotype calling, de novo RNA-editing site identification and annotation for any species with matching RNA-seq and DNA-seq data. In contrast to REDItools, it is equipped with statistical models (Bayesian and Binomial) to infer the reliability of homozygous genotypes derived from DNA-seq data and to distinguish RNA-editing sites from sequencing errors by assigning a p-value to each RNA-editing candidate. It runs faster than REDItools and is designed to run multiple samples in parallel.

We also recommend DARNED database (https://darned.ucc.ie) – a collection of annotated C-to-U and A-to-I editing sites in human, mouse and fly.

The processing of BS-Seq data and sequencing data after APOBEC treatment is challenging due to deamination reaction, which reduces significantly sequence complexity. Because conversion occurs only at C/5mC and not at G, strands are not complementary. These are the main concerns that computational tools must deal with, which is different compared with tools for regular DNA sequencing. A high quality of sequencing reads is crucial to get a good alignment and to correct methylation scores. Raw reads should be first trimmed using for example Cutadapt or Trim Galore software (https://github.com/FelixKrueger/TrimGalore) to remove adapter sequences and low-quality bases from the 3′end. Incorrectly converted reads should be discarded. Next, trimmed sequencing reads are aligned to the reference genome. To do this, two types of algorithms are available: wild-card or three-letter. The wild-card algorithm substitutes Cs with Ys in the reference genome, so reads can be aligned with both, Cs and Ts (tools: LAST, BSMAP, RRBSMAP, and Pash). The three-letter algorithm converts all Cs into Ts, both in the reference genome and in the reads (tools: Bismark, BRAT-BW, BS Seeker 3). Due to the asymmetrical alignments and non-complementarity, post-alignment tools (such as BSPAT and SAAP-RRBS ) are also useful. The aligned reads can be subjected to postprocessing quality control steps. E.g. PCR duplicates should be discarded, end-repair errors should be reduced, and reads that contain excessive cytosines in non-CpG context should be removed to reduce nonconversion errors. The remaining good quality alignments can be used for cytosine methylation and hydroxymethylation calling by e.g. Bismark methylation extractor.

• Bismark is a software package for the time-efficient analysis of BS-Seq data which performs both read mapping and methylation calling. Bismark is available under the GNU GPLv3+ licence at www.bioinformatics.bbsrc.ac.uk/projects/bismark/. The output of Bismark is easy to interpret and is intended to be analysed directly by the researcher performing the experiment.