Comparing exome sequencing data

[sadiqsaleem09]

Hi folks!

I recently got my exome sequencing raw data from macrogen. Along with providing the raw data, they have also provided an analysis file in the excel format listing all the SNPs with all the relevant information. I have few questions regarding those files.

1. How do I filter through the excel file and find the relevant mutations?
2. For each sample I have an excel file for germline DNA and for tumor DNA. How do I compare the two and find out true somatic mutations?

I am a newbie, so ain't good with handling raw data. However, in the mean time, I want to see if meaningful information can be extracted from the excel files.

I will really appreciate any input from the community.

Thanks,

Sadiq

[sadiqsaleem09]

Just to add to the above post:

Each excel file has the following column for each SNV:

chr_name
chr_start
chr_end
ref_base
alt_base
hom_het
snp_quality
tot_depth
alt_depth
region
gene
change
annotation
dbSNP135_full
dbSNP135_common
1000G_2010Nov_allele_freq
1000G_2011Oct_allele_freq
SCS : Clinical Significance
CLN : Variant is Clinical(LSDB,OMIM,TPA,Diagnostic)
OMIM : Variant has OMIM ID

[Bukowski]

Hi folks!

I recently got my exome sequencing raw data from macrogen. Along with providing the raw data, they have also provided an analysis file in the excel format listing all the SNPs with all the relevant information. I have few questions regarding those files.

1. How do I filter through the excel file and find the relevant mutations?
2. For each sample I have an excel file for germline DNA and for tumor DNA. How do I compare the two and find out true somatic mutations?

For point 1) Relevant to what exactly?
For point 2) I'd go back and ask for the VCF files and then use standard tools such as bedtools or vcf-tools to do the comparisons

Given that you appear to be doing tumour/normal comparisons were the genotype calls done with software that is designed to handle the biases in tumour cells? (VarScan, SomaticSniper spring to mind).

If you have the bam files, or mpileup output then you might want to look at approaches specifically designed for interrogating tumour/normal pairs (SomaticSniper, jointSNVmix).

I'm not really sure Excel files are the best way of handling and interrogating exome data.

[sadiqsaleem09]

Thanks Bukowski!!

Of course Excel is not the best format to deal with but since the other files are so huge to get uploaded, I was thinking if I can analyse these excel files in the meantime.

1) By relevant, I meant if the SNVs are true mutations or not

The information manual provided to me by the company mentioned that SamTools were used to detect SNPs and Indels. So I am assuming they must have used pileup and varFilter for that purpose.

I have the fastq, bam and bai files for each sample (germline/tumor pair).

I will check out the tools that you have mentioned to specifically interrogate the pairs. How are these tools different from bedools or vcf-tools?

[sadiqsaleem09]

Thanks Bukowski!!

Of course Excel is not the best format to deal with but since the other files are so huge to get uploaded, I was thinking if I can analyse these excel files in the meantime.

1) By relevant, I meant if the SNVs are true mutations or not

The information manual provided to me by the company mentioned that SamTools were used to detect SNPs and Indels. So I am assuming they must have used pileup and varFilter for that purpose.

I have the fastq, bam and bai files for each sample (germline/tumor pair).

I will check out the tools that you have mentioned to specifically interrogate the pairs. How are these tools different from bedools or vcf-tools?

[swbarnes2]

If you are looking to get the intersecting, or non-intersecting SNPs between control and normal, vcftools or BEDtools can do that. The good news is that your file is fairly close to bed format already, so you could try bedtools intersectBed

[sadiqsaleem09]

Thank you swbarnes2..

Are there any Windows based open-source platforms/softwares to use these tools?
Thanks

[sadiqsaleem09]

Thank you swbarnes2..

Are there any Windows based open-source platforms/softwares to use these tools?
Thanks

[Bukowski]

Thanks Bukowski!!

Of course Excel is not the best format to deal with but since the other files are so huge to get uploaded, I was thinking if I can analyse these excel files in the meantime.

1) By relevant, I meant if the SNVs are true mutations or not

Define 'true' I mean the question is still very open, a 'true' mutation I would think is one that you have gone back and verified via another genotyping mechanism.

'Genuine somatic mutations' would be verifiable changes between your tumour/normal pairs

'Causative mutations' would be somatic changes presumably driving the tumour phenotype.

Assuming you're talking about the latter (again I'm a bioinformatician not a biologist) I guess you're looking for mutations in known cancer genes - perhaps compare against a database such as COSMIC?