Biologists have long sought to understand the impacts of deleterious genetic variation on fitness and population viability. However, our understanding of these effects in the wild is incomplete, in part due to the rarity of sufficient genetic and demographic data needed to measure their impact. The genomics revolution is promising a potential solution by predicting the fitness effects of deleterious genetic variants (genetic load) bioinformatically from genome sequences alone, bypassing the need for costly demographic data. After a historical perspective on the theoretical and empirical basis of our understanding of the dynamics and fitness effects of deleterious genetic variation, we evaluate the potential for these new genomic measures of genetic load to predict population viability. We argue that current genomic analyses alone cannot reliably predict the effects of deleterious genetic variation on population growth, because these depend on demographic, ecological, and genetic parameters that need more than just genome sequence data to be measured. Thus, while purely genomic analyses of genetic load promise to improve our understanding of the composition of the genetic load, they are currently of little use for evaluating population viability. Demographic data and ecological context remain crucial to our understanding of the consequences of deleterious genetic variation for population fitness. However, when combined with such demographic and ecological data, genomic information can offer important insights into genetic variation and inbreeding that are crucial for conservation decision making.

RNA sequencing (RNA-Seq) is a popular method for measuring gene expression in non-model organisms, including wild populations. While RNA-Seq can measure gene expression variation among wild-caught individuals and can yield important biological insights into organism function, sampling methods may also influence gene expression estimates. We examined the influence of multiple technical variables on estimated gene expression in a non-model fish, the westslope cutthroat trout (Oncorhynchus clarkii lewisi), using two RNA-Seq library types: 3’ RNA-Seq and whole mRNA-Seq. We evaluated effects of dip netting versus electrofishing, and of harvesting tissue immediately versus 5 minutes after euthanasia on estimated gene expression in blood, gill, and muscle. We detected 30% more genes with whole mRNA-Seq than with 3’ RNA-Seq and found that 58% of genes were significantly differently expressed between 3’ RNA-Seq and whole mRNA-Seq. Our findings indicate that 3’ RNA-Seq and whole mRNA-Seq are robust to the technical variables related to the field sampling approaches tested here with a lack of differential gene expression among sampling methods and tissue collection time after euthanasia. However, we found that gene expression varied based on which RNA-Seq library type was used on the same set of samples. Our study suggests researchers could safely rely on different fish sampling strategies in the field and save money and analyze more individuals using 3’ RNA-Seq, but should use whole mRNA-Seq when working with a species without good genomic resources, and when maximizing the number of genes identified and detecting alternative splicing are important.

RNA sequencing (RNA-Seq) is becoming a popular method for measuring gene expression in non-model organisms, including wild populations sampled in the field. While RNA-Seq can be used to measure gene expression variation among wild-caught individuals and can yield important biological insights into organismal function, technical variables may also influence gene expression estimates. We examined the influence of multiple technical variables on estimated gene expression in a non-model fish species, the westslope cutthroat trout (Oncorhynchus clarkii lewisi), using two RNA-Seq methods: 3’ RNA-Seq and whole mRNA-Seq. We evaluated the effects of dip netting versus electrofishing, and of harvesting tissue immediately versus 5 minutes after euthanasia on estimated gene expression in blood, gill, muscle, and liver. We found higher RNA degradation in the liver compared to the other tissues. There were fewer expressed genes in blood compared to gill and muscle. We found no difference in gene expression among sampling methods or due to a delay in tissue collection. However, we detected fewer genes with 3’ RNA-Seq than with whole mRNA-Seq and found statistically significant differences in gene expression between 3’ RNA-Seq and whole mRNA-Seq. The magnitude and direction of these differences does not appear to be dependent on gene type or length. Our findings indicate that RNA-Seq is robust to the technical variables related to the field sampling techniques tested here but varies based on the tissue sampled and the RNA-Seq library used. This study advances understanding of usefulness of RNA-Seq to study gene expression variation in evolution, ecology, and conservation.