I am putting this question because I did not find any useful information from internet because of limited access. My question is related to control (or normal) data that we use for somatic mutation detection (with the help of mutect or SomaticSnipper etc) with tumor data.

I don't have control data and my aim is to find somatic mutations in Multiple myeloma WES caner data.

Here are the questions:

Can we use Hg19 as a control data? If not, what are the technical reasons of not using hg19 as a control data because it hg19 is consensus call.

Is there any way to infer (or predict) somatic mutations (either statistically or by something else) out of all possible mutations? For this reason, I want to know what information is given in control data.

  • $\begingroup$ Could you explain what articles or what research have you done to answer this questions. Also this is a list of questions, could you focus on one and open another post with others (ie question 5 is unrelated to question 1-4), so ask it in other posts (but be aware that you should show your research in each question you post) $\endgroup$
    – llrs
    Apr 6, 2018 at 7:18
  • $\begingroup$ @Llopis: Thanks for suggestion. I tried to add more descriptions and make it more valuable. $\endgroup$ Apr 6, 2018 at 9:22

1 Answer 1


Without control data from your subjects, I don't think there's really no way to distinguish somatic mutations from germ-line mutations. The best you can do is to screen out common variants, which are germ-line mutations that are shared by large numbers of individuals using the population frequencies from something like the Exome Aggregation Consortium:

annotate_variation.pl -downdb -webfrom annovar -build hg19 exac03 humandb/
annotate_variation.pl -filter -build hg19 -dbtype exac03 example/ex1.avinput humandb/
  • $\begingroup$ @heatthobrien: Thanks for the reply. What about hg19 reference genome as a control data. Can I use it as a control data for all patients? $\endgroup$ Apr 17, 2018 at 5:21
  • 1
    $\begingroup$ hg19 is a composite of a small number of individuals, but it doesn't capture any of the variation between individuals. You can use it as a reference to identify variants, but most of the variants you find will have been inherited from you're subjects' parents. $\endgroup$ Apr 17, 2018 at 5:37
  • $\begingroup$ My idea was to use probability mapping and frequency mapping from a consensus call to indentify somatic mutations with the help of motif visualization tools. It may be little bit difficult, time taking or prone to error method. But I have here no choice because I don't have the control data. I also don't know if it is right way. (But sounds practical because we can statistically infer that most probabely it can be somatic mutation. $\endgroup$ Apr 17, 2018 at 7:21
  • $\begingroup$ One more suggestion can you tell me is there any way to use Genome Aggregation Database in ANNOVAR because I have checked the ANNOVAR doccumentation and I don't find any way to use this database. $\endgroup$ Apr 17, 2018 at 7:23
  • 1
    $\begingroup$ Without control data any method you try will be a waste of time. $\endgroup$
    – Devon Ryan
    Apr 17, 2018 at 8:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.