The prediction of non-coding RNA (ncRNA) expression, structure and function is a rapidly expanding field of research. A great variety of ncRNAs with different regulatory, catalytic and structural functions have been described. We performed in silico experiments to predict and classify ncRNAs of the protozoan Trypanosoma cruzi, the causative agent of Chagas disease. 4195 ncRNA candidates were identified through comparative genomics between T. cruzi and Trypanosoma brucei using eQRNA to identify compensatory mutations. From the 1382 candidates which did not present significant protein-coding potential, 49 were classified as tRNAs or rRNAs and 29 showed similarity to previously characterised ncRNAs from public databases. Here, we describe a novel in silico protocol for the identification of ncRNAs in different life-cycle stages of T. cruzi. We have compared the mapping efficiency of over 22 million publicly available RNAseq reads of the Y strain to each of the 8 currently sequenced T. cruzi genomes using BWA and Bowtie. The best results regarding mapping quality were obtained with Bowtie allowing 3 mismatches between aligned sequences and against the genomes of CL Brener Esmeraldo-like, CL Brener non-Esmeraldo-like and Sylvio strains. To account for problematic data features, such as short length, the genetic difference between strains, and the poor assembly and annotation of the genomes, we decided to only consider those reads which mapped to orthologous regions in all the three aforementioned genomes. The final sets of ncRNA candidates from both strategies were compared and further annotated based on currently available ncRNA databases. These were then submitted to structural analyses using the DotAligner algorithm for RNA structure clustering. Next we intend to assess the different functional classes of ncRNAs from T. cruzi and contribute to a more thorough understanding of the role of these RNAs in parasite evolution, development and pathogenicity.