There are two ways: Assembly by mapping and de Novo assembly.
If you already have a human genome file, you can take each DNA piece and map it to its closest match in the genome. If you can cover the whole genome this way, you are done.
The alternative way is to exploit overlaps between DNA fragments. If two 1000 bp pieces overlap with 900 basepairs, that's probably because they come from two 1000 regions of your genome that overlap by 900 baswpairs. You can then merge the pieces. By iteratively merging millions of fragments you can reconstruct the original genome.
Both these approaches are surprisingly and delightfully deep computational problems that have been researched for decades.
If you already have a human genome file, you can take each DNA piece and map it to its closest match in the genome. If you can cover the whole genome this way, you are done.
The alternative way is to exploit overlaps between DNA fragments. If two 1000 bp pieces overlap with 900 basepairs, that's probably because they come from two 1000 regions of your genome that overlap by 900 baswpairs. You can then merge the pieces. By iteratively merging millions of fragments you can reconstruct the original genome.
Both these approaches are surprisingly and delightfully deep computational problems that have been researched for decades.