Knockout alleles are powerful tools for functional gene analysis during homeostatic and pathological conditions, but they frequently have their use restricted due to pleiotropic effects and embryonic lethality that arise as a consequence of inactivation of a specific gene. Conditional gene knockouts emerged as a way of overcoming these limitations by promoting gene modification at different levels: tissue, cell lineage and developmental period. The Cre/lox is the most used technology in conditional gene studies and the development of the Tol2 transposon transgenesis system boosted the application of this method, especially through the generation of numerous Cre driver lines. However, limitations of random and multiple-copy integrations associated with Tol2 transgenesis demonstrated to affect the accuracy of spatiotemporal analyses. Recently, the rise of genome engineering technologies, such as CRISPR/Cas9, paved the way for the development of precise targeted knock in strategies that allowed the creation of superior Cre- and lox-containing alleles, but they still lack efficiency and reproducibility. Based on this gap, my colleagues and I established an approach called GeneWeld for promoting precise targeted integration of diverse cargo DNA at high frequencies using short homology-mediated end joining. We demonstrated that homology stretches as short as 24 or 48 base pairs and a universal CRISPR gRNA to induce exposure of these homologous sequences in vivo, combined to our donor vector series, pGTag (plasmids for Gene Tagging), are sufficient to promote gene targeting in zebrafish, pig, and human cells. Fluorescent reporter and Gal4 cassette targeting frequencies were up to 10-fold higher than similar previous work and, overall, the average transmission rate was of 50% across different loci. The straightforwardness and efficiency of GeneWeld led to the development of more sophisticated tools, specifically conditional alleles. First, I generated neural lineage specific Cre drivers by integration of a Cre recombinase cargo at endogenous transcription factor genes in zebrafish. By targeting ascl1b, olig2 and neurod1 loci, which exhibit neural lineage expression, the goal was to use their native promoters to drive the recombinase expression. For this, I used our pPRISM-Cre vector that in addition to the Cre cassette, it has a secondary marker for allele tracking. I observed considerable rates of germline transmission (5%-33%), and all the three lines presented Cre expression correctly matching the endogenous patterns of the target genes. Next, I used our UFlip (Universal Flip) vectors for generating dual-function revertible conditional alleles at rbbp4 and hdac1 loci. The UFlip vectors consist of a rox/lox-flanked cassette containing a gene trap and a secondary marker. Intronic targeting of the UFlip construct in passive or active orientation leads to "gene on" and "gene off" alleles, respectively. Using this strategy, I was able to isolate rbbp4-gene On and hdac1-gene Off alleles with 9% and 10% efficiencies, respectively. Injecting Cre RNA into 1-cell stage embryos, I demonstrated effective recombination at the lox sites for both alleles, which confirmed the functionality of the gene trap cassette through disruption of gene expression and observation of mutant phenotypes. Our UFlip alleles represent a breakthrough, since they are the first reported revertible alleles in zebrafish, allowing conditional experiments at a higher level not ever performed in this model system. Combining the Cre and the UFlip lines we are going to be able to analyze neurogenesis, brain tumorigenesis and regeneration processes by looking directly to neural stem, progenitor and post-mitotic cells. More important than that, the work presented in this dissertation can be easily extended to any type of conditional study and it has the potential to bring great progress to the field through the generation of more powerful tools.