There are research groups working on this exact question; understanding cold adaptation in enzymes. You can read about this in several articles such as 'Computation of enzyme cold adaptation', or 'Molecular Structural Basis for the Cold Adaptedness of the Psychrophilic β-Glucosidase BglU in Micrococcus antarcticus', or 'Specific amino acids responsible for the cold adaptedness of Micrococcus antarcticus β-glucosidase BglU'
On a general level there are several sructural characteristics that are typically found. They are usually far more flexible (i.e. less salt-bridges and other such stabilising bonds). The flexibility come from having fewer strong bond between and within the polypeptide chain, and even containing disordered regions. The flexibility also means that they are usually far less stable, and more difficult to work with. They are industrially interesting, as developing enzymes that can work at low temperatures can reduce the cost of running reactions. This is essentially a step towards a greener-industry. There are also speculations about various ratios and abundance of the amino acid types, and where they are located. You would have to read articles on specific proteins to investigate this.
From a practical view-point, cold adapted enzymes are usually isolated from arctic microorganisms and characterised in the lab - in order to determine their temperature tolerance. The structure of such enzymes are then solved (after lots of work) and form the basis for understanding cold-adaptation.
To answer the question more specifically, no its not realistic to know if something is coldadapted directly by looking at the protein sequence (on its own), this has to be determined experimentally in the lab. However, if the protein is from a cold adapted organism (typically arctic bacteria) there is reasons to assume it is cold adapted before you test it in the lab. You can also apply som bioinformatics (e.g. blast search) to determine if its related to cold adapted proteins. Also, If that specific type of protein that you are looking at, is well studies, you could also try to make homology models and see if its structurally similar to cold-adapted relatives or if it is more structurally similar to its mesophilic or even thermophilic counterparts. This is however very speculative, and i would not put too much weight on theoretically assumed adaptations - you should at minimum know that the protein of interest is from a cold adapted organisms, to even begin to look (computationally) on why and how it could be cold adapted.
