Script to handle Sanger sequencing data automatically
-
Python 2.7+
-
Have MAFFT command line tools installed . Can be obtained at http://mafft.cbrc.jp/alignment/software/
-
Python Libraries :
- Pandas (+ Numpy) 0.20.1+
- BioPython 1.69+
- adamP_BioTools v.1.16+
-
Sequencing data. Either ABI files, or FASTQ files. FASTA can of course also be aligned, but has a loss in precision due to the absence of PHRED scores.
-
Reference Sequence
Download to your project directory from GitHub, or clone it on the command line using the following command:
git clone git://github.com/adamnicolaspelletier/sequencing-alignment-sanger.git
Sanger sequencing is used very often in molecular biology. Yet, unless your desired sequence is short, if you try to align your forward and reverse runs against one-another, you do not have 100% of your sequence with good confidence. Sometimes, you will have inconsistencies/conflicting results (indels, SNVs) between the 2 sequencing reactions - what should you believe?
Biologists tend to handle this manually in this case: visualize the sequences with software and look at the phred scores. And compare with the reference sequence to see what they should expect. But this is not good enough for high-throughput approaches: manually curating hundreds of sequencing runs is not only mind-numbing, but also more and more error-prone as time goes.
There are free common tools that perform alignments very well, such as BLAST, but it only performs Local Alignment over one sequence, while Sanger Sequencing of larger constructs can benefit from Multiple Alignment of both Forwad and Reverse runs against the reference. Of course, this assumes you know the identity of the sequenced DNA. If you do not, a quick BLASTn will give you the highest probable hit. Moreover, BLAST does not handle quality scores from FASTQ files, which makes it harder to determine a consensus when there is a discrepancy between 2 sequences.
Other free options (e.g. ClustalOmega) can perform multiple sequence alignment , yet none offer high-throughput capability.(isoform alignment, variant alignment, quality scores handling for discriminating).
This is a simple approach to take FASTQ files and do global alignment using MAFFT command-line tools. When supplied with reference sequence for different isoforms, the script performs multiple alignment and returns the alignment in CLUSTAL format, asks for the user input to select the appropriate isoform, then generates a clustal output and fasta file. (sequencingalignment.py)
If necessary, the user can also supply fasta sequences with containing variants, which is useful to detect if a construct has acquired a desirable mutation after site-directed mutagenesis.
This script was designed with high-throughput in mind: meaning if you only need one or 2 sequences aligned, you may find you are jumping through unnecessary hoops. However, it remains possible to use with this in mind.
The example files included in the docs directory show how data is meant to be organized for the script to function properly.
The flags have default values that use the example files included in the docs directory, so the script can be used without any flags. However, you will need to use the flags if you wish to use other files.
python sequencingalignment.py -fq path/to/my/fastqfile.fq -f path/to/my/fastqinfofile.txt -r path/to/my/reference/sequences.fa
If you have more than one reference sequence (VERY LARGE FASTA file, with hundreds of entries and likely irrelevant sequences), the script will find the entries with the same Ensembl ID as your sequencing reaction form the filename file you supplied .
However, if one wants to determine which isoform his amplified sequence corresponds to, thus more than 1 reference sequence, the script will show the alignment with all of them, letting the user choose which isoform he wants to align to.
The script also includes a variant mode, which lets the user visualize his alignments with variant(SNV, mutations, indels) fasta sequences of the reference sequence. This is useful after cloning a potentially mutated gene, or inducing directly mutagenesis. Variant FASTA sequences can be obtained from sitedirmutagen.py script, from the same author.
python sequencingalignment.py -vm -v path/to/my/variantFASTA.fa
Finally, the script has optional FASTA clipping functionality to clean out output and facilitate visualization, by removing unnecessary bases outside of the reference sequence.
python sequencingalignment.py -c
Use the -h flag for further options and documentation
- Add functionality for multiple FORWARD and REVERSE contigs for a given reference sequence at once. Useful for long genes that require multiple reactions
to span the full length. - Add support for regional phred scoring comparison, instead of per base. Will judge whether upon encounter of a discrepancy, whether the base call is a in less reliable region of the sequencing reaction, such as often found in the extremities.
- Add optional automatic isoform detection, instead of user choice. Optimal for high-thoughput alignment.
- Add optional testing of sequencing run: sometimes Sanger sequencing is bad, or basecalling is wrong, and a high-thoughput approach needs to be able to identify this automatically.
IPYTHON has problems with the argparse module, other Python distributions do not. I suggest using Official Python in the meantime. Script has not been tested in Python 3.
Please open an issue for support.
Please contribute using Github Flow. Create a branch, add commits, and open a pull request.