Bioinformatics textbooks explain one method to reconstruct the whole DNA
sequence from multiple reads. The method builds a graph and finds a Eulerian path in it. The Eulerian path corresponds to a DNA sequence.
However, the graph can contain more than one path. How do we know which path corresponds to the actual sequence?
If the DNA sequence contains repetative elements, and your reads are not long enough to resolve them, you won't know if you have the right path. Usually you split your results into contiguous regions whose sequence you are confident in.