Looking for mutations in PacBio cancer data: an alignment-free method
MICADo is a tool to perform variant calling on targeted (third) next-generation sequencing data. It's algorithm is based on colored de Bruijn graphs. It has been designed for high-precision variant calling for each sample in a cohort. The intuition behind the method is the substraction of systematic biases or errors present in a cohort in order to single out real mutations. We evaluated its precision on targeted sequencing data generated by a PacBio sequencer.
For any information or help running MICADo, you can get in touch with:
MICADo installation requires an Unix environment with python 2.7, it is therefore Linux and MacOS compatible. You will need a compatibility layer (bash shell) for running Linux applications on Windows.
MICADo was implemented in Python (python 2.7; http://www.python.org/) and tested under CentOS 6.6, Linux (Ubuntu 14.04 and 16.04) and Mac OS environments.
git clone https://github.com/cbib/MICADo.git
Python modules required are listed in requirement.txt. Use the following command line to install them:
cd MICADo
sudo pip install -r requirements.txt
MICADo runs in a command line environment. The package contains one main script called MICADo.py that coordinates the execution of the whole process. This section describes how the user must call it.
- Fastq file of the sample of interest, targeted capture.
- Fastq files of the samples of the cohort, generated using the same library preparation and the same sequencer.
- Reference sequence of the targeted region (FASTA). This can be a multi-fasta file describing e.g. multiple isoforms.
- (Optionnal) A TSV file describing known SNPs.
/!\ Reads in fastq files must be oriented in accord to the reference fasta file. You can do it with /bin/orient_reads_in_forward_direction.py or with the cutadapt tool (http://cutadapt.readthedocs.io/en/stable/index.html).
/!\ Make sure to copy the whole cohort fastq files in the folder data/fastq/
python2.7 src/MICADo.py
--samplekey <sample_name> # Sample label for the results
--fastq data/fastq/<experiment_name>/<sample_name.fastq> # Sample fastq file (with all the path)
--experiment <experiment_name> # Experiment label corresponding to the folder name containing fastq files of the whole cohort
--fasta data/reference/<reference.fasta> # Reference sequence in fasta format
[options]
--snp SNP file for reference sequence (file name, with all the path)
--min_support_percentage Minimum of read support percentage for node filter (default=3.0)
--npermutations Number of permutations / random samples to perform (default=1000)
--max_len Maximum allowed indel length (default=250)
--pvalue p-value threshold for significance (default=0.001)
--results Output results as JSON file (file name, with all the path)
--kmer_length Size of k-mer words (default=20)
--disable_cycle_breaking Do not search for k-mer values yielding a DAG
Note that k-mer length increases automatically to obtain a DAG (Directed Acyclic Graph) (stop at k >= 70). Use the --disable_cycle_breaking to allow cycle(s). In this case, this part of the graph is not analysed.
Use http://www.snp-nexus.org/ to generate a refseq TXT file for known SNPs containing SNP description on reference transcript.
Then use /bin/make_SNP_file.py to construct the final TSV file (example for the NM_004119 transcript of FLT3 gene):
python2.7 /bin/make_SNP_file.py
NM_004119.2.fasta # Corresponding fasta file
refseq_22221.txt # Refseq file obtain from snp-nexus
snp_FLT3.tab # File name generated from make_SNP_file.py
NM_004119 # Corresponding transcript
200 # Number of nucleotides before and after SNP
See /data/reference/snp_FLT3.tab for the final format.
Usually, results and log are displayed in the terminal. You can also write results and log in separate tsv/log files with:
python2.7 src/MICADo.py
--samplekey <sample_name>
--fastq data/fastq/<experiment_name>/<sample_name.fastq>
--experiment <experiment_name>
--fasta data/reference/<reference.fasta>
> <sample_name>.tsv 2> <sample_name>.log
If you specified a SNP file, a TSV file will be created (and the direction /output/SNP/) with, for each SNP, the following informations: sample key, SNP ID, the number of reads supporting the reference nucleotide and the number of reads supporting SNP.
See http://www.json.org/ for complete description of json file format.
Example from the TP53 cohort (see associated paper). The pre-built library used for sampling is provided.
python2.7 src/MICADo.py
--samplekey C_256_1
--fastq data/fastq/TP53_C/C_256_1.fastq
--fasta data/reference/NM_000546.5.fasta
--snp data/reference/snp_TP53.tab
--experiment TP53_C
--kmer_length 18
--results C_256_1.json &> C_256_1.log
In both C_256_1.log and C_256_1.json files, we can see that the only alteration with p-value < 0.001 and z-score > 10 is C > T at position 866 of NM_000546.5 which is the expected mutation for this sample.
- v.1.0, 2015-12-01, first release, version used in the accompanying paper
Copyright (c) 2015
Justine Rudewicz (1,2) ([email protected])
Hayssam Soueidan (1) ([email protected])
Macha Nikolski (1,2) ([email protected])
(1) CBiB - Universite Victor Segalen Bordeaux,
146, rue Leo Saignat, 33076 Bordeaux, France
(2) CNRS / LaBRI, Universite Bordeaux 1, 351 cours de la Liberation,
33405 Talence Cedex, France
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.