Single nucleotide polymorphisms are the most abundant polymorphisms in the genomes analyzed to date. They are becoming the main choice of molecular markers for breeding, genotyping, and diagnosis purposes, due to the large amount of sequences data available. Identification of those nucleotide polymorphisms will provide useful markers for genetic mapping, population genetics and association studies. It will also provide criteria to infer the evolutionary history of the analyzed genes, which can be relevant to select the best candidate genes to test in future association studies. For those reasons, the objectives of this work were: 1) identify and validate both in silico and in vivo, the SNPs and INDELS existing in EST resources; and 2) analyze the nucleotide diversity in Coffea spp., in addition to selected C. arabica cultivars. A Pipeline for identification of SNPs and INDELS was generated using sequences from the Brazilian ESTs Coffee Genome Project as well as other Coffea sequences available in GenBank. The pipeline was carried out by a haplotype -based strategy to detect reliable SNPs in 23.019 contigs assembled. A total of 23.062 SNPs e 2.165 INDELS were identified in 5184 contigs with more than four ESTs assembled. With the haplotype-based strategy, it was possible to define the probable ancestral of C. arabica transcripts. The majority of ESTs from C. arabica, came from only two different alleles, providing molecular evidences about C. arabica speciation. According to our analysis, approximately 55% of C. arabica sequences were derived from C. eugenioides, and 45% were considered as come from C. canephora. Interestedly, C. eugenioides contributes mostly with genes related to basal metabolism and the secondary metabolism, while genes C. canephora genes are involved with signal transduction and gene expression regulation. The in vivo analyses are being performed by sequencing PCR fragments of several genes in 24 Coffea genotypes corresponding respectively to 12