In the C. elegans germline, multiple mechanisms act to silence foreign DNA. Because these pathways can silence experimentally introduced transgenes, they pose a technical challenge to biologists studying the germline and early embryo. Developing strategies to overcome or bypass silencing of transgenes would facilitate research in multiple fields that use C. elegans embryos as a model system. One pathway that can silence single-copy foreign sequences involves recognition of the foreign DNA by genomically encoded 21U-RNAs, which initiate silencing through the Piwi Argonaute PRG-1 (Shirayama et al. 2012). The diversity of 21U-RNAs is high enough to target essentially any DNA sequence. Endogenous germline genes are protected from silencing by a second pathway that involves sequence-dependent licensing by 22G-RNAs bound to the Argonaute CSR-1 (Seth et al. 2013). Newly generated transgenes can adopt either an expressed or a silent state as a result of competition between these two pathways. In some cases, recognition of the GFP portion of a transgene by the PRG-1 pathway can lead to complete silencing. Because silencing is sequence-dependent, we thought it should be possible to design protein coding sequences that evade recognition by PRG-1 and thereby escape silencing. Designing sequences that lack 21U-RNA recognition sites is difficult because the number of 21U-RNAs is very large, and their base pairing specificity (i.e., mismatch tolerance) has not been well characterized. However, we reasoned that bona fide germline-expressed genes should be depleted of high-affinity 21U-RNA binding sites, and might also be enriched for 22G-RNA binding sites that could facilitate expression. Therefore, we developed an algorithm that constructs a coding sequence for any protein of interest by assembling short “words” found in germline-expressed genes. We refer to coding sequences generated in this way as “germline optimized.” We constructed synthetic transgenes to test whether germline-optimized sequences are resistant to silencing. We obtained robust germline expression of bacterial proteins that were otherwise efficiently silenced. Moreover, by using a germline-optimized GFP coding sequence, we were able to obtain expression of gfp::cdk-1, a model transgene that is particularly prone to silencing. We have produced a web-based tool that allows users to evaluate the degree of germline optimality of existing transgenes and to design germline-optimized coding sequences for any protein of interest.