If you are interested in a ssDNA that folds on itself instead of making a double-helix, you should look for tools aiming at modeling RNA. Some of them where developed for RNA but include alternative parameters for ssDNA. In particular, you can have a look at the Vienna Package (https://www.tbi.univie.ac.at/RNA/#). Typically, you would first predict the most probable 2D-structures, then model corresponding 3D-structures. If your DNA sequence is quite long, you may want to model in 3D only parts of the 2D-structures (those of interest, and/or those with highest confidence on the 2D prediction).

No - you will have to go step-by-step:

First determine your open reading frames and translate your nucleotide sequences to a protein sequences. If your single stranded DNA is a viral DNA, it may encode more than one protein, and you may even have overlapping coding sequences in different reading frames. You may also need to look more closely at the biology of your virus, since some viruses encode long open reading frames, where the individual mature proteins are the results of proteolytic processing.

Once you have figured out the individual protein sequences, you can use these for a blast search on the pdb. Look at the individual hits to see whether the structures can not only serve as templates for the individual protein chains, but also help you to build the supermolecular assemblies.

The best method to model the individual protein structure depends on the level of similarity between the query sequence and best template structures: For high similarity, you can use traditional homology modeling (e.g. MODELLER). For intermediate similarity, a threading algorithm (ITASSER) will work best. For very low homology, an ab-initio modeling method may be needed, but this won't be easy and may not yield good results.

Be aware that errors made in the early steps will be amplified in subsequent steps!

Annemarie Honegger Isaure Chauvot deBeauchêne The sequence is for ssDNA aptamer, which folds itself to a three dimensional than to double-helical. I have performed a long time (1 us) simulation from the linear chain and it is folded. Is it an acceptable structure or is there any method to validate it?

Your MD-derived structure should conform to the predicted secondary structure

MFOLD/UNAFOLD may be useful

http://unafold.rna.albany.edu/?q=mfold/DNA-Folding-Form

also see

https://www.nature.com/articles/s41598-017-01348-5

https://www.cell.com/biophysj/comments/S0006-3495(15)02975-6