I have a raw NGS data for a photosynthetic microorganism and I am willing to do an SNP detection analysis on this data. What kind of downstream treatments should be done to raw NGS data to prepare it for SNP detection? Can anyone tell me which tools should I use? Thanks in advance.
The process could be something like:
1) Read processing
2) Read mapping
3) SNP calling.
Usually the raw read data needs to be processed to remove the low quality reads that could introduce noise in the SNP detection. There are some tools to do that, but the most used are FastX-toolkit and Fastq-mcf from ea-utils (the first one is available in Galaxy if you want to use a GUI). Suggested parameters a min. qscore of 30.
For the read mapping there are several options too, personally I prefer Bowtie2, BWA or Stampy. Once you have the mapping file (sam/bam) you can filter it and get rid of the unmapped reads using samtools (samtools view -F 4).
Finally the important step, the SNP calling. I don't recommend samtools. Take a look to a good post where they compare different tools (http://bcbio.wordpress.com/2013/10/21/updated-comparison-of-variant-detection-methods-ensemble-freebayes-and-minimal-bam-preparation-pipelines/). Two recommendations GATK and FreeBayes. GATK already have a page of calling snps best practices that you could take a look (https://www.broadinstitute.org/gatk/guide/best-practices#dnaseq-ovw-ovw).
I hope this helps,
The process could be something like:
1) Read processing
2) Read mapping
3) SNP calling.
Usually the raw read data needs to be processed to remove the low quality reads that could introduce noise in the SNP detection. There are some tools to do that, but the most used are FastX-toolkit and Fastq-mcf from ea-utils (the first one is available in Galaxy if you want to use a GUI). Suggested parameters a min. qscore of 30.
For the read mapping there are several options too, personally I prefer Bowtie2, BWA or Stampy. Once you have the mapping file (sam/bam) you can filter it and get rid of the unmapped reads using samtools (samtools view -F 4).
Finally the important step, the SNP calling. I don't recommend samtools. Take a look to a good post where they compare different tools (http://bcbio.wordpress.com/2013/10/21/updated-comparison-of-variant-detection-methods-ensemble-freebayes-and-minimal-bam-preparation-pipelines/). Two recommendations GATK and FreeBayes. GATK already have a page of calling snps best practices that you could take a look (https://www.broadinstitute.org/gatk/guide/best-practices#dnaseq-ovw-ovw).
I hope this helps,
Hi,
I think the post from Aureliano Bombarely pretty well covers it. I would like to add the following to this.
Regarding 1) Read processing: Sometimes, trimming of possibly remaining adapters is also necessary. For tools that do this, you might want to have a look at Trimmomatic and alternatives. One should actually not underestimate preprocessing. There is quite some truth to the saying "Garbage in, garbage out"...
Regarding 3) Maybe you want to have a look at https://twitter.com/bioinformer/status/515151684121923585/photo/1 -> It is already a bit older (O'Rawe et al., 2013) but illustrates nicely the variation in variant calling with different tools. Accordingly, to start with, you might want to be rather strict and take the *overlap" of different SNP/variant callers.
Best,
Cedric
I agree with Aureliano and Cedric. In addition, you may include steps like: 1) Only picking the unique alignments from the alignment (bam) files and 2) InDel re-alignment of the bam files as additional steps. I too suggest to use either GATK/FreeBayes. Thanks!
Thank you so much for your answers. Your comments gave me a good workflow to follow.
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioscience
- Biotechnology