Just to expand on Pauls answer a little you only start obtaining decent sequence 30-40 bp from the end of your primer.  To properly check sequence you really need confirmation of the sequence in both directions and as such when sequencing a PCR product this is not possible.  Cloning the product enables you to prime sequencing reactions from the plasmid sequence flanking the insert enabling full sequence coverage of your insert.

pcr sequencing is quick and easy .PCR introduces random errors and sequencing the product will give the consensus sequence which is usually correct in that minority peaks are not visible. Cloning takes more time but has the advantage that all sequences can be sequenced with the same plasmid primers.The downside is that each pcr product is now clonal so if you choose a minority clone it will sequence perfectly but the sequence may not be representative of the true sequence if it contains a taq induced error. Cloning also allows you to grow up large amounts of product when you have identified the correct clonal sequence

Just to expand on Pauls answer a little you only start obtaining decent sequence 30-40 bp from the end of your primer.  To properly check sequence you really need confirmation of the sequence in both directions and as such when sequencing a PCR product this is not possible.  Cloning the product enables you to prime sequencing reactions from the plasmid sequence flanking the insert enabling full sequence coverage of your insert.

I found that PCR products can be sequenced double stranded when you add an adapter sequence to the 5' end of the primer. Adapter sequences such as M13. That way you don't have to use PCR product specific oligos for sequencing but can use universal oligos.

E.g.

5'TGTAAAACGACGGCCAGT-(here your spec. forw. sequence)

5'CAGGAAACAGCTATGAC- (here your spec. reverse sequ.)

I also found it that cleaning up the PCR product with columns often results in better templates for sequencing than an ExoSap clean up step.

I would recommend cloning if you expect a mix of sequences.

After reading Natascha's answer I had a look back at some of my sequencing and have to confess my off-the-cuff (and somewhat dated!!) 30-40 bp was a little on the conservative side and completely agree that actually perfectly good sequence can be obtained 20-25 bp from the end of the primer.

Thank you everyone for your very interesting answers. Actually, my aim is to study the possible polymorphism (but not at the individual level, more at the sex level or tissue level) in mRNA (after reverse transcription into cDNA) sequences encoding for a secreted protein. I have found some companies offering the double strand sequencing directly from cleaned PCR products as said by Natasha. I'm also using a Pfu polymerase to limit PCR random errors. Also I was not to worried about the flawed sequences obtained from primers end since I'm using primers out of the CDS (in the sequence encoding for the peptide signal and in the polyA tail). So I was more encline to work from PCR products since I'm more interested in the consensus sequence from a pool of cDNA representative of a group of samples (same sex and/or same tissue) and I was afraid of misinterpreting putative sequence variation from minor clones as esxplained by Paul... But I'm still hesitating looking towards the publication since I'm "afraid" that future reviewers could judge this strategy as flawed and reproach us for not cloning sequences...  May I ask you all what is your opinion about it?

Thank you very much for your help!

Expanding a little bit more on the previous answers...

First to sequence a PCR product, as you should envisage from Paul's answer, you need to have a "clean" PCR product, which means an amplification without extra non-specific bands (or at least they should be in such minority that you won't detect them on an agarose gel), otherwise your sequencing will be hard to read. Still, you can purify the fragment to sequence from a gel if your amplification is not specific.

Secondly, Companies don't spend too much time on cleaning up the sequencing reaction, so you end up with what Marcus points out, that the non-incorporated fluorescently labelled dideoxy-nucleotides exiting the column in a large Gaussian cover the signal from the shortest sequencing fragments up to a length of 30-40 nt (the first ca. 20 being your primer). But if you run the reaction with the thermosequenase (whether also the sequencing does not matter) in-house, it is very simple to precipitate the sequencing reaction with 2 M NH4OAc and 2.5 vol EtOH... You'll get a very clean sequencing trace from the first nucleotide after your primer.

Good luck.

Hello

I have done both types, sequencing before and after cloning. I recommend the after cloning method generally and especially when you need a complete sequence.

good luck 

Regarding your aims of 'studying possible polymorphisms' are you looking at abundance? functional studies? distribution?  Either way you will need to accurately identify the SNPs.  Sequencing pools of PCR fragments (regardless of how clean they are) in the first instance may, I think, cause some issues regarding a) missing low abundance polymorphism b) necessarily following up on false positives.  Personally, in the first instance I don't think there is any substitute for the full sequencing of the GOI in 30-40 individuals (hopefully your gene isn't too long).  Once the SNPs are identified you can then design allelic discrimination assays for broader screens or generate clones as required.

Hii

I am doing both the ways for sequencing. Its depends on the individual convenience and availability of time and resources. I completely agree with the above answers given by Paul and Marcus. To specifically Cecile question, the answer is no. No reviewers can judge direct PCR product sequencing strategy as flawed, until your polymorphic base is coming in the start of end of the sequences. As an example, You might have seen bar-coding papers 1000s of samples analyzed with a particular region. That will be done through direct pcr product target sequencing, and I was following the direct PCR sequencing for all my bar-coding studies. I would recommend cloning PCR fragments for identifying novel genes and cross validating it.

So in my opinion, you can go ahead with the PCR product sequencing and I hope no reviewers will ask questions.

Good Luck