What are the differences between asTair v2 and v3?

There are some changes in default settings of the asTair modification calling between versions 2 and 3, and some updates of the tool features have taken place, resulting in the following differences:

1) maximum read depth used for the modification ratio calculation | 8000 asTair v2 | 250 asTair v3 | 2) lower base quality cutoff | 13 asTair v2 | 20 asTair v3 |

3) output and data manipulation by cytosine context | No asTair v2 | Yes asTair v3 |

4) working with both single-end and pair-end input data | No asTair v2 | Yes asTair v3 |

5) GZIP compression of table outputs | No asTair v2 | Yes asTair v3 |

6) The depth (DEPTH column in v2 and TOTAL_DEPTH in v3) in the .mods file is calculated differently:

| Mod + Unmod asTair v2 | all bases on both strands passing quality filters asTair v3 |

NB: At present, an * is assigned as a modification value at a position where no reads cover the strand of interest even if there are reads in the total depth column. This value was 0 for asTair versions <= 3.26, and is customisable with the --no_information parameter.

What is the most efficient way to run asTair?

We suggest using the --per_chromosome option whenever possible and to split per context with the --context options for faster genomic analyses. Calling modifications in --context CpG mode would require less than 10GB of RAM for large mammalian genomes, whereas the --context all mode could need more than 30GB of memory.

My laboratory uses the NCNN 160mer spike-in from your original publication TET-Assisted Pyridine Borane (TAPS) sequencing. How can I analyse the data?

The NCNN 160mer used as a spike-in in our original paper requires the user to split reads by cytosine context at known modified positions and could have significant coverage. For these reasons, we created an extra script that can be found in https://bitbucket.org/snippets/gvelikova/7eGMqg:

git clone https://gvelikova@bitbucket.org/snippets/gvelikova/7eGMqg/160mer-ncnn-and-highly-covered-short.git

python 160mer-ncnn-and-highly-covered-short/separator.py -i NCNN.bam --read_length 75 --modified_positions 51,111 --modified_positions_orientation OT,OB -d output_dir/ 

Can you call modifications on scTAPS with asTair?

At present, astair call with the --library reverse option could give information about modifications under certain scTAPS experimental designs. scTAPS analysis features are under development.

Installing asTair without pip

If pip is not available for some reason, you can simply download the package directly and install it manually:

wget https://bitbucket.org/bsblabludwig/astair/get/master.tar.gz
# or if you don't have wget, try
# curl -O https://bitbucket.org/bsblabludwig/astair/get/master.tar.gz

tar -xzf master.tar.gz -C astair_3 --strip-components=1

cd astair_3

We would recommend using a virtual environment to avoid issues with globally installed packages, or if you are on a shared system and need to install locally:

virtualenv venv

# Activate the environment
source venv/bin/activate

To install, execute:

python setup.py install
# or, if you want to install as a user, but without using a virtual environment:
# python setup.py install --user

Now, you can perform all analysis and processing steps described above.

If you were using a virtual environment, you can deactivate the environment after you're finished using astair by calling:

deactivate

Cite asTair

Working on asTair was part of my doctoral thesis, and can be cited as:

Velikova, G. V. (2020). Development and application of computational methods to study DNA modifications [PhD thesis]. University of Oxford.