Hi Sina,

there is something I really don't understand, probably because of my complete ignorance on how transgenic Zea mays actually works. Given that your target sequence is transgenic (i.e. exogenous), how is it possible that different SNPs are found within that sequence? I assume that the transgene was generated from a single, CLONAL sequence (e.g. a plasmid containing the target gene). If this is correct you should only expect a single SNV to exist for each position of the transgene (even if in homozygous state or in case of endoreplication, only exception being somatic mut). Am I missing something?

Besides from the previous consideration, did you try to map the target sequence on the reference Zea genome? although unlikely, is there any chance that a similar sequence is found in ref? did you try to visualize the bam/bai alignment files using IGV ?

Final point: do you know the position of the transgene in the ref genome? if so it could be interesting to setup a PCR with a primer mapping in the ref in close proximity to the transgene and one in the transgene itself, right after the SNP to assess if with this approach 'heterozygosity' disappear.

All the best,

Rocco

is it possible that your sequencing is picking up the change in mitochondria where heteroplasmy is often found or in chloroplast sequences ?

Dear Rocco,

you are right. Normally, only  a single SNV should be possible. My only explanation for my findings are somatic mutations. And I am not so familar with endoreplication. I thought endoreplication can also introduce mutations. But maybe I am wrong. Does anyone have another explanation for two SNV? This is problem no. 1 of my findings. Problem no. 2 is that the detected SNP at Position 100 has very different frequencies in different samples.

At the moment I am analyzing the methylation status of the affected position because methylation may be the reason for the C to T substitution.

There is no similar sequence in the maize genome. My explanation above was not 100% right. The forward primer of my analyzed target region is located in the transgene and the reverse primer is located in the natural maize genome. I have analyzed the 3´end of the transgene. 

Dear Paul,

I cannot exclude your idea. My analyzed maize seeds are commercial and normally the transgene should not be in mitochondria or chloroplasts. The company Claims that the transgene is only  integrated in the normal Genome with one copy.  

However, do you have an idea how to test this?

Sorry Sina plants are not really my area but if you can do a dna extraction that selects for mitochondrial dna at the expense of genomic dna and the ratio of the bases changes from that found by using your dna extraction method that might be an indication

Hi Paul, hi Sina,

from my understanding the new information about primer design (primer 1 in the transgene, primer 2 in maize genome)  rules out the (appealing) hypothesis that transgene could be integrated in mitochondria. Do you agree?

Hi Rocco, Sina

I think that you are correct that it is in the genome if the genome primer is a good primer annealing close to its optimum Ta. My suggestion only works if the pcr conditions are allowing the reverse primer to anneal to an only  similar sequence at too low a temperature giving a pcr product. Clearly this is unlikely to be the same size as the genomic amplimer so yes..I do agree with you which is a bit of a shame because it is hard explaining the different snp ratios unless both targets exist in the genome but one primer is just on the edge of working so one amplimer gives varying amounts of product  between samples but really I am struggling for a reasonable rationale here. I suppose there is one thing I have to check and that is the state of the negative control sample. Hopefully it is negative and that post pcr contamination from previous experiments  ( low level) is not an issue here

Hallo Rocco and Paul,

many thanks for your helpful thoughts. You are right, mitochondria is unlikely because of my primers. My negative controls were always negative. Post PCR contamination is also unlikely but I can not exclude this 100%. Maybe I have to extract some DNA samples again (I have still a small rest of every maize grain) and then sequence them.

And what do you think of my idea that somatic mutations are the cause. And that the SNP position is a strong methylated C which becomes then a T. This process may happen to different percentages, what explains the different SNP ratios. Or do I have a thinking error?

Hi Sina,

your hypothesis could definitely make sense, explaining the reason for the strange ratios between the two nucleotides. However this implies that the specific nucleotide is a terrific  mutational hotspot. Explaining such a huge error rate on that position is challenging based only on spontaneous deamination of methylated cytosine..consider also that deamination would end-up in a mismatch between the two strands, triggering repair mechanisms. Do you know if the specific variant you detected is a known polymorphism in the genome of the transgene ?

As far as I know, we are the first group who detected this SNP. But there are not many people in the world working on this topic.

Hi Sina, I mean in the genome of the species from which the transgene was cloned (i.e. not maize).

Hi Sina,

It's an interesting question. Firstly, are you really really sure you are using the right method to detect your SNPs? (I don't know which is the best but I know different methods will give you different ratios)

Secondly, I'm not sure I understood your experimental setup, you say each sample is from a different grain, and that you are also studying different varieties. Does that mean each sample is also from a different plant, or do you have some samples that are more related than others. Could you compare related samples to see if they all have the same mutation rates? Are all these varieties related or were they made by independent transformation events? Have you tried to sequence a sample from leaves or stem to check if it's seed-specific (gene methylations can be tissue-specific)?

Are you absolutely sure it is only present in one copy in the genome? Could it have been duplicated onto the sister chromatid? Could it have been tandem duplicated? Could it have been "corrected" by the internal machinery by using a similar gene as a template? I'm not sure if Zea does endoreduplication in it's seeds... the literature would surely have information on that.

I hope you find the answer. Best,

Anna

Dear Sina, 

But what is your reference or control? Do you compare sequenced fragment to the transgene sequence + original maize variety? Are you sure that IDENTICAL transgene sequence integrated into the genome? How the transgenesis was performed, by a plasmid - than did you sequenced also the plasmid as the reference? Was the vector homogeneous at this region? So, did you received reliable information from the manufacturer?

So to my point of view, you should have at least three sequence references: 1 - the transgene itself, eg from the manufacturer; 2 - transgene from the vector; 3 - sequenced region where the integration happens from maize.

As you wrote, all DNA is done from a single grain - what was their physiological stare? if you separate the embrio from the rest of the grain will you obtain similar results? As you mention endoreplication it might be an issue, as the process is more probable in the endosperm. Besides, why you would expect just a single as the genome is polyploid and polyploidity increases once with endoreduplication? 

One more - did you checked the number of transgenes per genome? whould it correlate with the number /quantity of vectors used to consider transgenesis successful?

So, I think Rocco is right with his arguments. Besides, mitochondrial genome is very unlikely to be the target as otherwise you would obtain different non-transgenic region of the sequence.

Good luck!

First, you didn't say what pipeline are you using, as well as which filters are you applying. The first thing I can suggest is to filter by depth and quality of of the SNPs calling. Also I suggest to get only biallelic sites and keep only homozygous and heterozygous SNPs.

If you didn't use the GATKs pipeline you should give it a shot.

The variation of the alleles at the particular segregating sites among your 50 samples could suggest that that is a very variable locus. On the other hand, did you removed PCR duplicated? or, did you have more than 1 exon in your amplicon? Or, what is the length of it?

Best,

Erik

Dear Rocco,

sorry for my late reaction. My variation is not in the genome of the species from which the transgene was cloned. The affected base is an A in this species and after transformation into maize it is a C and my findings are C and T. 

Dear Anna,

I am sure that my method is very appropiate for my purpose.

I have ~ 100 to 200 single grains of different varieties. The varieties are not related. Even the grains of the same variety do not have similar mutation rates. But all varieties harbor the same transformation event. I did not sequence other tissues. But you are right, it would be interesting. Sure, only one copy in the genome.

Hi Sina,

that's very puzzling: is it technically possible for you to FISH the transgene in your target plant? If not you could try a quantitative-PCR on gDNA using your standard primers (i.e. one in one out) and a second set (two-in, possibly one identical and similar amplicon length) to assess if the second pair gives a much stronger signal.

Dear Levitchi,

thanks for your comments. I do not have all the reference sequences you mentioned. I compare it to the sequence which is present after transformation. This sequence is verified by the company. And there is only one copy in the genome. This is also verified by the company. I think they are reliable.

It would be interesting to analyze embryos alone. But it is a challenge to seperate the embryo from the rest very clean. I am afraid to contaminate the embryo with endosperm.

Dear Erik,

I´ve analyzed the samples with MiSeq from Illumina. And for the evaluation (pairing, trimming, quality filter, variant calling) I have used CLC Genomics Workbench from Qiagen.

My amplicon ts 180 bp. I think I do not understand what you mean with "removing PCR duplicate"? Only one exon..

Thank you!

Hi Rocco,

I can not perform FISH. But real-time PCR. Can you explain your suggestion more in detail. I  do not understand the point.

Hi Sina,

sorry for late answer. I was just considering the possibility of your transgene being present in multiple copies. So, simplifying a bit: are you 100% sure about the specificity of the reverse primer mapping to the maize genome ? no chance it falls into a repeat region? did you try to align it to the ref maze genome to assess its mapping specificity?

Hi Rocco,

the reverse primer can also bind in the mitochondrial genome.

I like your real-time PCR approach, but I am afraid of a signal bias. If I chose the same primer pair and for the comparison the same fwd but another reverse primer, then the second amplicon is much shorter. I think this will influence the fluorescence signal. Maybe it is better to design a second new primer pair at the 3´end of the transgene with similar size.

Hi Sina,

the reverse primer can also bind to mitochondria?!!

That's a very important piece of information, as it brings back into life Paul Rutland's hypothesis of mitochondria integration. So the very first thing to do should be to change the reverse primer to get rid of this abnormal binding and repeat at least a few Sanger seq.

Dear Rocco,

several scientists tested the copy number of this transgene using different methods, digital PCR, real-time PCR and Southern Blot. All scientists came to the same result: Only one copy. But I will test another reverse primer to be 100% sure.

Dear Sina,

did they already test the presence of the transgene in mitochondria? did they use exactly the same reverse primer you're using in your assay? If they used a different, gDNA-specific reverse primer they might have completely missed this information. In brief, it is just a matter of a PCR + seq: I would give it a shot.

Hi Rocco,

not exactly in mitochondria. And not with the same reverse primer. You are right, maybe they have overlooked it. I will keep you up to date;-) Many thanks for your help.

Dear Rocco,

I have performed Sanger Seq because it is easier for me. I detected always 2 Peaks at the SNP position in my mutated samples. I have used another fwd and another reverse primer now and there is only one peak at this position. Really, this is crazy. I have to perform NGS to be really sure. But wow, this is a little sensation. Now, we have to determine the locus of the transgene fragment plus the mitochondrial part. I have read, that mitochondrial DNA is often inserted in the nuclear DNA in plants. It is an ongoing process. And often there are multiple copies of mitochondrial DNA in the genome (hundreds or thousands). These mitochondrial DNA copies play a role in double strand break repair. They are often located directly near double strand breaks and something similar may happen through genetic engineering. Thank you very much Rocco!

Hi Sina, that's good news, at least there is a trail to follow!

Fingers crossed for your next exps!  :)

Well done Sina and Rocco it is an interesting problem and worked out well. It would be fascinating to hear how it turns out  exactly when you have fully cracked it Sina.

all the best,Paul