Which strategy or strategies should I use?

[qiuyuxinhan]

I am a newer for NGS. Recently I got 8 samples of NGS data from the Solid 5500xl. The average depth is about 10x, some maybe deeper. All of the samples are sporadic. People who has this kind of disease will die before they grow up. So lots of evidences make us believe that this kind of disease should be a de novo mutation. Which strategy or strategies should I use that could help me find the right gene? Thank you very much!

[dpryan]

Are the samples from related individuals?

[Zaag]

Is it whole genome, exome, panels? Do you have reads or alignments or variants?

[qiuyuxinhan]

They don't have any relationship with each other.

[qiuyuxinhan]

We did the whole exon sequence by the solid 5500xl.

[swbarnes2]

In brief, align to reference genomes, human, I guess, then call SNPs. Best case, you are looking for a gene where all your samples have SNPs which cause highly deleterious changes to the protein. If your samples are not related, they probably won't all share the exact same mutation.

If you don't have that, you'll have to think a little harder. Maybe there are a few genes mutated in the same pathway? Maybe you have a CNV? It's always possible that your mutation is not even in the exome, in which case, your data won't show it.

[qiuyuxinhan]

First of all, thank you very much for your guidance! We have checked the mutations which all 8 samples shared. But we didn't find any reasonable gene for our case. We also checked the CNV before we did the NGS. As you said, maybe there are a few genes mutated in the same pathway, but how could we annotate this? Which software could we use to analysis this ? Thank you again!

[Brian Bushnell]

Let's assume this is a one-or-two-gene disorder, since otherwise, most methodologies just won't work.

It's unreasonable to expect unrelated people to share a mutation. More likely, the related affecteds have a homozygous mutation in the same genes(s), due to inbreeding, which is highly probable for fatal genetic conditions. In this case it is best to sequence related individuals and find genes in which the unaffecteds have heterozygous mutations, but the affecteds have homozygous mutations.

One way to isolate responsible genes is to sequence multiple generations of related individuals (like great-grandparents, grandparent, parents/aunts/uncles, siblings/cousins, etc), and use software to determine inheritance. You don't even need sequencing for most of them; a cheaper SNP array is fine. This will allow you to isolate the genes (or areas) that are unique from the rest of the family (due to crossover events) and thus may be causing problems. Then, armed with the most probable locations, you can examine the affecteds' sequences in detail to determine the problem.