Dear Emilian & Gael, All the shared problems can be solved by using physical mapping technology, from OpGEN. OpGEN compliments NGS data & help in refining the data, filling the gaps, align the contigs more accurately etc. It just help you in getting excellent data.
From what little I know I believe different types of next gen sequencing result in different problems. Roche 454 Pyrosequencing seems to be bad with indels, but gives you nice long reads. Illumina Solexa seems to be good with indels, but bad with short sequence repeats and have high substitution errors due to modified polymerase and reverse transcriptase. If you look at ABI-SOLiD it is easier to tell between a sequencing error vs a polymorphism, but again you have the problems associated with a short sequence read.