In past years, papers have tried to identify the subset of all genes in a genome that are essential for viability. In “The Indispensable Genome,” Science Magazine considers this capability a turning point in biology:

Game-changing moments in functional genomics often reflect the development and application of powerful new reagents and methods to provide new phenotypic insight on a global scale. Three independent studies describe systematic, genome-scale approaches to defining human genes that are indispensable for viability, which collectively form the essential gene set. On pages 1092 and 1096 of this issue, Blomen et al. (1) and Wang et al. (2), respectively, report a consistent set of ~2000 genes that are indispensable for viability in human cells. Moreover, very similar results were obtained by Hart et al. (3). For the first time, we now have a firm handle on the core set of essential genes that are required for human cell division. This opens the door to studying the roles of essential genes, how gene essentiality depends on genetic and tissue contexts, and how essential genes evolve. [Emphasis added.]

This achievement follows on the heels of yeast studies where researchers found only 1/6 of its 6,000 genes to be essential. That’s an astonishingly low fraction. What functions do essential genes perform?

The yeast essential genes encode proteins that drive basic cellular functions such as transcription, translation, DNA replication, cell division cycle control, and fundamental metabolism. Moreover, the yeast essential genes share several attributes that reflect their critical role in cellular life. For example, they are often conserved and evolutionarily constrained, are highly expressed, and encode abundant proteins that tend to form stable complexes and thus are rich in protein-protein interactions.

The landscape of essential genes in human cells can now be explored using the conceptual framework established in yeast.

All three studies on human cells found only 10 percent of the 20,000 genes in the human genome are essential.

What Is the Other 90 Percent Doing?

Boone and Andrews, authors of the review, indicate that patterns in the human essential genome are similar to those in the yeast essential genome:

All three groups found that human essential genes are highly conserved, and much like yeast, they encode abundant proteins that engage in protein-protein interactions. The core set of human cell essential genes also tend not to be duplicated and appear to have increased evolutionary constraints, as they evolve slowly and are associated with fewer deleterious single-nucleotide polymorphisms. Although many essential genes are involved in fundamental biological processes including transcription, translation, and DNA replication, a substantial fraction remains functionally uncharacterized. Indeed, each analysis prioritized a wealth of uncharacterized genes whose essential roles are waiting to be explored.

That leaves about 18,000 “non-essential” genes to explore. One possibility is that they really are essential, too, when partnered with other genes. Yeast cells, for instance, can survive two non-lethal single mutations, but die when both occur together. This is called “synthetic lethality.” Researchers have identified hundreds of thousands of these synthetic lethal interactions in yeast, they say. Initial studies in humans show the following pattern (notice the use of the phrase “functional information”):

Blomen et al. begin to address the extent of synthetic lethal interactions in human cells by screening a set of five nonessential genes with roles in secretion for synthetic lethal negative genetic interactions. They discovered an average of ?20 synthetic lethal double-mutant interactions for a given nonessential gene, and these interactions tend to occur with functionally related genes. Even this relatively small genetic network suggests that the properties of the extensive genetic networks mapped for yeast are conserved and can now be mapped efficiently in human cells. The genetic network described by Blomen et al. ought to catalyze large-scale, collaborative efforts to map genetic interactions in human cells. Such an effort promises to enable functional annotation of the human genome, because genetic interaction profiles are rich in functional information and provide a quantitative measure of gene function.

These discoveries have the effect of raising the number of essential genes. If a cell can’t survive a double hit on two interacting genes (a “synthetic lethal” condition), this indicates functionality even if each gene can take a hit on its own.

The studies of Blomen et al., Wang et al., and Hart et al. reveal the core essential gene set for human cells, setting the stage for the next wave of new genetic and chemical-genetic science that will take place directly in human cells. A future challenge will be to develop genetic tools, such as conditional alleles of essential genes, for exploring the terminal phenotypes and the various molecular mechanisms underlying the lethality associated with perturbation of different essential functions.

Life Will Go On

How often do double mutants occur in non-essential genes? If infrequent, synthetic lethals will be invisible to purifying selection. A non-essential gene can mutate and life will go on. Wang et al. say as much:

Essential genes should be under strong purifying selection and should thus show greater evolutionary constraint than that of nonessential genes. Consistent with this expectation, the essential genes found in our screens were more broadly retained across species, showed higher levels of conservation between closely related species, and contain fewer inactivating polymorphisms within the human species, as compared with their dispensable counterparts (Fig. 2, E to G). Essential genes also tend to have higher expression and encode proteins that engage in more protein-protein interactions.

Blomen et al. claim similar findings: essential genes show more conservation. They claim that “old” essential genes “emerged in premetazoans.” But then, to their surprise, they found new essential genes incorporated into old existing functional genes:

Remarkably, the products of “new” essential genes are more often connected with old rather than other new essential gene products, suggesting that they largely function within ancient molecular machineries (fig. S9, B and C).

In PNAS, Rubin et al. looked for “The essential gene set of a photosynthetic organism.” They identified “718 putative essential genes” for the photosynthetic lifestyle of a cyanobacterium. 

There are certain limitations to the essentiality information determined here. Although we identified genes that are essential to the organism when individually mutated, they do not represent a minimal gene set. Essential processes for which there are redundant genes will not be discovered using an approach based on single mutants. In S. elongates, however, this complication is of lesser concern than in most other cyanobacteria because of its small genome size, which at a streamlined 2.7 Mbp, harbors little redundancy. In addition, the findings of essentiality reported here apply only to the specific laboratory conditions used and are likely to be different for a subset of genes under other growth conditions. Finally, because ncRNAs, regulatory regions, and other intergenic regions are much smaller, on average, than protein-coding genes, the essentiality calls for these regions are inherently of lower confidence than those made for protein-coding genes. Therefore, conclusions of essentiality for non-coding loci and to a lesser extent, protein-coding genes must be validated by targeted mutation before definitive statements can be made about their essentiality.

In short, the count of essential genes will vary by lab and researcher. This is definitely a work in progress, so measures of essential genes will need refinement with more study.

Contrasting Predictions

Let’s integrate this information by contrasting the predictions of design and Darwinism about essentiality. The distinctions are not clear cut. On the one hand, Darwinians would like to see purifying selection acting to eliminate non-essential genes because, for large populations like yeast, they incur a metabolic cost. On the other hand, Darwinians have historically appealed to “junk DNA” and “vestigial organs” to explain things that do not appear essential, pointing to the weakness of purifying selection to eliminate mutations.

Design advocates, too, maintain competing expectations in tension. They would like to find functions for all gene activity to falsify the junk-DNA myth. But they would like to allow for functions beyond mere survival: functions that a designer with an artistic taste would create for beauty and pleasure.

So who’s winning this debate on essentiality? It’s too early to tell. Not enough is known yet. Based on experience with ENCODE and modENCODE, it seems likely that more functions will be found for everything in the genome (barring neutral or near-neutral mutations), but they will not always be essential for survival. There is a wealth of phenotypic evidence to support this: the beautiful spirals of a conch shell, elaborate patterns in fur and feathers, and other cases of elegant design that seem to go beyond the requirements for reproduction. Animals could satisfy Darwin’s criteria by being all grey and just getting by till they have offspring, but life is incredibly vibrant with “useless” beauty. One suspects that genotypic evidence will follow suit.

Boone and Andrews point to “a wealth of uncharacterized genes whose essential roles are waiting to be explored.” Who is better prepared to explain what the “substantial fraction” of genes that remain “functionally uncharacterized” do — those who start with the assumption that “if it works, it’s not happening by accident” or those who expect cobbled-together bits of junk?

Editor’s note: This article was originally published in 2015.