Is reproducibility of identified variants in a gene important in Sanger sequencing?

Question

[DanielC]

Dear Friends,

I have identified some known and novel variants in a gene by sanger sequencing. The identified variants are then used for protein-drug binding interaction studies (this is the main study of the research). Do you think, I should perform the reproducibility check on these identified variants? What if some variants are not reproduced, can I still do protein-drug binding interaction studies on them? Thanks

[grant.hovhannisyan]

better safe than sorry. Make a double check if you have a chance, it will pay off in a long run.

[ATpoint]

Is reproducibility important, well...make an educated guess... You can of course do research on irreproducible data but that is chasing ghosts, think about whether you want to invest time and effort (and motivation) into that.

[DanielC]

Thanks, but it does not answer the specific question asked.

[ATpoint]

It kind of does. If a variant is not reproducible it is most likely not real, why would you do any experiment on something that does not exist? You can of course do any in silico study on them, the question would be why.

[WouterDeCoster]

Probably because the specific question is based on common sense. Why would you start a bunch of other studies without first validating if the variant is really there?

[gb]

I think this is an important part:

protein-drug binding interaction studies

If you are researching something for the medical field I would say re-sequence and re-check it multiple times. Even if it is an early pilot, in further stages there are many fda rules.

What if some variants are not reproduced

Then you still don't know anything, you need minimal 3 samples or 3 checks. Do as many as you need to feel sure. You can also take a small (or really big) side road and try to find out why you sometimes find it and sometimes not.

can I still do protein-drug binding interaction studies on them?

This is still tricky, I think that if that variant has a good quality score and not something like double spikes you can trust that specific data from the sequencer. But what about errors in and early stage of the PCR. Do you still get that variant if you take 3 samples from the same subject (different days or different locations) or multiple subjects. I guess if you would find it only once it is maybe nothing.

Did you already checked if your gene of interest is present in other databases like genbank or something. You can try to find it and maybe use that first.

[GenoMax]

Do you think, I should perform the reproducibility check on these identified variants

Can you clarify if you sequenced a sample one time and went ahead with other studies? Are you referring to re-sequencing the same sample to make sure the sequence result is reproducible or other samples that exhibit same phenotype?

[DanielC]

Thanks.. Yes, sequenced a sample one time. Yes..referring to reseqencing same sample. Please let me know of your suggestions. Thanks

[GenoMax]

Considering the amount of effort you seem to have spent on this re-sequencing should be small potatoes. Use a couple of different primers to get different products while you are at it.

[DanielC]

Thanks.. After re-sequencing..if some mutations are not reproduced..do you think they can be studied in silico for protein drug interaction studies? Thanks

[ATpoint]

The consensus answer in this thread is no. Is it that you have already done the resequencing, cannot reproduce the variant but want to stick to your study because you already invested into it? Don't get me wrong, I fully understand the struggle you are probably in, but ask yourself if it is worth continuing to build a castle on sandy ground.

[GenoMax]

do you think they can be studied in silico for protein drug interaction studies?

As a theoretical exercise I suppose you could. If that mutation does not exist in nature there would be no practical application.

[DanielC]

Thanks for the answer genomax!