Question: What could be the reason for having mutations in too many cancer samples?
0
gravatar for MAPK
4.4 years ago by
MAPK1.5k
Washington University in St. Louis
MAPK1.5k wrote:

I have been introduced to genotypes data and have found some interesting driver mutations that were already discovered in cancer. However, I also have found a few mutations in some new genes. For one particular gene, I have found that mutation in almost 95% (n= 80) of cancer samples and not in controls. I was told that this could be somatic mutation, but I am finding it difficult to understand why it is somatic mutation. I would really appreciate if someone can clarify this riddle to me.     

ADD COMMENTlink modified 4.4 years ago by Manu Prestat4.0k • written 4.4 years ago by MAPK1.5k

Do you find the exact same somatic InDel in all patients? Even if that is not the case, 95% of samples having a somatic nonsynonymous/frameshift mutation in the same gene is very unlikely (see PMID: 26404825, Figure 2). Are the cancer patients related?

ADD REPLYlink written 4.4 years ago by trausch1.5k
4
gravatar for DG
4.4 years ago by
DG7.1k
DG7.1k wrote:

If you are doing sequencing in tumours you will be identifying both germline and somatic mutations. Depending on your allele frequency cutoffs and filtering criteria before you look at the variants, you may be looking mostly at somatic variants because tumours are constantly acquiring new mutations. If you are doing matched tumour/normal sequencing then you will be filtering out the majority of germline variants and looking only at somatic mutations. Somatic mutations just means all mutations found in somatic (as opposed to germline) cells, so all mutations that an individual has acquired during their lifetime, mostly we study these in tumours but if you picked any random skin cell from a person you would likely find some somatic mutations due to DNA damage, chemical exposure, etc.

Now, the question is whether you mean that that particular gene has some mutations (but not necessarily the same mutations) in 95% of cancer samples and not controls, or you are finding the same recurrent mutation? Are you sequencing tumours or are these germline samples from people with some hereditary cancer syndrome? What sort of filtering are you doing in order to remove common variants?

ADD COMMENTlink modified 5 months ago by RamRS27k • written 4.4 years ago by DG7.1k

Thank you Dan for your answer. No, all these samples contain exactly the same mutation (an indel) and is also not in homopolymer region as Manu hinted above. These are tumour samples subtracted germlines; and to filter out common variants, we are using maf cut off of 1% (dbSNP).

ADD REPLYlink modified 5 months ago by RamRS27k • written 4.4 years ago by MAPK1.5k
1

Are these from a particular cancer type? Was there any selection going on with the patients? There are lots of known recurrent mutations in some cancers. FLT3-ITD, NPM1 indels, EGFR exon 21 indels, etc. It would seem unusual to find the same indel in 95% of samples unless there was something particular about the patients/tumours themselves. Like if you're selecting tumours that were all resistant to a particular therapy, or susceptible to a particular therapy, etc.

ADD REPLYlink written 4.4 years ago by DG7.1k

These samples are of mixed individuals of different age groups and yes they are of same cancer type. As you mentioned, "Somatic mutations just means all mutations found in somatic (as opposed to germline) cells, so all mutations that an individual has acquired during their lifetime". Now my question is how could all those individual get the same somatic mutation in the exact genomic region?

ADD REPLYlink modified 5 months ago by RamRS27k • written 4.4 years ago by MAPK1.5k
1

Like I said, there are lots of known recurrent mutations found in cancer. BRAF V600E for instance is found in lots of different cancers. The exact same mutation. Seeing a mutation in lots of different tumours either within a sub-type of cancer, or across different cancers, is one of the factors that goes into defining a mutation as a driver mutation. Remember that tumours are a) heterogeneous b) prone to mutation and c) undergoing an evolutionary process including selection and competition. So you see recurrent mutations for a few different reasons. One is because some genes are prone to mutation by various mechansisms (trinucleotide repeats for instance) meaning some genes and regions within a gene are just easier to mutate. We also see recurrent mutation because although the mutation is acquired somewhat randomly, there is a filtering process going on. Lots of mutations confer chemo resistance or growth advantages to tumour cells.

ADD REPLYlink modified 5 months ago by RamRS27k • written 4.4 years ago by DG7.1k
1

If you are going to continue doing work in genomics of cancer I would highly recommend going through some of the literature or books about cancer genetics and mechanisms, etc. As a bioinformatician never forget that the biology knowledge is first and foremost. You can't analyze data if you don't understand the biology that underlies it.

You may have something interesting on your hands, but the sheer number of samples with the mutation is somewhat suspicious, even for the same cancer sub-type without some other sort of patient or tumour related criteria that went into sample selection (like therapy response). Unless this is a rare sub-type for instance that hasn't been looked at in depth genomically yet. Always think about possible sources of error or less exciting/interesting biological reasons for the observation :)

ADD REPLYlink written 4.4 years ago by DG7.1k
1

Thanks a lot, this is very helpful information. Sometimes digging into literature is too-much-information to digest, especially when you are transitioning from different background. But again, I have found your answers very helpful to understand this.

ADD REPLYlink modified 5 months ago by RamRS27k • written 4.4 years ago by MAPK1.5k

Glad to help with a bit of a crash course. Finding some good reviews or well written books is the key to getting started with the literature when making the switch.

ADD REPLYlink written 4.4 years ago by DG7.1k
1
gravatar for Anima Mundi
4.4 years ago by
Anima Mundi2.7k
Italy
Anima Mundi2.7k wrote:

Hello,

a somatic mutation is simply a mutation occurring in a somatic cell. Somatic cells are all those cells (the vast majority in our bodies) that do not belong to the germinal line (the one originating gametes). Germinal mutations are especially meaningful in biology because they can be transmitted to the progeny, and they are the driving force of evolution. Unless your tumor derives from the germ line, any new mutation you observe will be a somatic one. Of course, one can be born with some alleles exposing him to the risk of cancer: the mutation originating that allele would be a germinal one, the following mutation causing the onset of cancer (and others that might follow) would all be somatic.

So, variations that are present in all cells of a person are most likely inherited, those found in a percentage of cells from the same person are likely to be somatic mutations.

ADD COMMENTlink written 4.4 years ago by Anima Mundi2.7k
1
gravatar for Manu Prestat
4.4 years ago by
Manu Prestat4.0k
Lyon, France
Manu Prestat4.0k wrote:

Are you sure this is not a sequencing error or bioinformatics pipeline issue? Check if this mutation occurs in an homopolymeric region, or within a di/tri nucleotides stretch or next to a primer end, those are often sources of False Positives.

ADD COMMENTlink written 4.4 years ago by Manu Prestat4.0k

No this is not in homopolymer region. I had a few of those, different mutations in the same homopolymer region and I did not count them as "interesting variants".It is actually an indel. Thanks for your suggestion.

ADD REPLYlink modified 4.4 years ago • written 4.4 years ago by MAPK1.5k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 759 users visited in the last hour