DEG 15, an update of the Database of Essential Genes that includes built-in analysis tools

1Essential genes refer to genes required for a cell or an organism to survive under certain conditions. The research on the determination of essential genes has attracted significant attention in the past decade, due to its theoretical implications and practical uses. Studies of genome-wide gene essentiality screenings have elucidated fundamental cellular processes that sustain life. We created DEG, a Database of Essential Genes in 2003, a time when the genome-scale gene essentiality screening was still not available. The development of DEG parallels with the development of the essential-gene field. Significant progress has been made in performing genome-wide essentiality screenings among diverse species, primarily due to technological developments. We subsequently published DEG 5, which included essential genes of both bacteria and eukaryotes, and DEG 10, which included both protein-coding genes and non-coding genomic elements. Since 2014, when DEG 10 was published, significant progress has been made mainly owing to the invention of CRISPR/Cas9 and the widespread use of Tn-seq. To accommodate the progress in essential-gene studies, we created DEG 15, which, compared to DEG 10, provides two major updates:

2The number of essential-gene entries has significantly increased. Specifically, compared to DEG 10, the number of eukaryotic essential genes has increased by more than fourfold, and that of prokaryotic ones has more than doubled. Of note, the human essential-gene number has increased by more than ten-fold. Figure 1 shows the number of essential gene records in different versions of DEG, as well as the methods used to determine the gene essentiality. It is shown that the increase in prokaryotic records and eukaryotic records are mainly due to the widespread use of Tn-seq and CRISPR/Cas9, respectively (Figure 1).

3We have developed built-in analysis modules by which users can perform various analyses, such as essential-gene distributions between bacterial leading and lagging strands, sub-cellular localization distribution, gene ontology and KEGG pathway enrichment analysis, and generation of Venn diagrams to compare and contrast gene sets between experiments.

4Genome-wide essentiality screenings have elucidated the molecular underpinnings of many biological processes in prokaryotes. However, limited knowledge has been gained regarding essential genes in human cells. Large-scale gene essentiality screenings across human cell types can reveal genes that encode factors for regulating tissue-specific cellular processes, and such screenings in cancer cells can disclose factors that determine cancer phenotypes, thus revealing important targets for cancer therapies. However, genome-wide inactivation of genes in human cells and the analysis of lethal phenotypes have been hampered by technical barriers.

5One of the major breakthroughs in biotechnology has been the invention of CRISPR/Cas9 (CRISPR-associated RNA-guided endonuclease Cas9), which is a simple yet powerful tool for editing genomes. Cas9, an endonuclease, can be guided to specific locations within complex genomes by a guide RNA (gRNA). Cas9-mediated gene editing is simple and scalable, enabling the examination of gene functions at the systems level. Because of the ease and efficient targeting, CRISPR/Cas9 is described as being analogous to the ‘search’ function in a modern word processor. The invention of CRISPR/Cas9 has revolutionized the biological research in many fields, with essentiality screenings in human cells being no exception.

6In 2015, three papers were published simultaneously reporting the genome-wide identification of essential genes among diverse human cell types. Wang et al. used the CRISPR-based approaches in analyzing multiple cell lines, and found tumor-specific dependencies on particular genes. The core-essential genes among these cell lines are enriched for genes with evolutionarily conserved pathways, with high expression levels, and with few detrimental polymorphisms in the human population. Analysis by Blomen et al. revealed a synthetic lethality map in human cells. Hart et al. used CRISPR-based approaches to screen for fitness genes among five cell lines, and consequently discovered 1580 human core fitness genes, and context-dependent fitness genes, that is, genes conferring pathway-specific genetic vulnerabilities in cancer cells.

7The technology of CRISPR can be used on various cell types. Mair et al. used the CRISPR system to catalogue essential genes that are indispensable for human pluripotent stem cell fitness. Lu et al. determined genes essential for podocyte cytoskeletons based on single-cell RNA sequencing. Wang et al. used CRISPR in identifying essential oncogenes for hepatocellular carcinoma tumor growth. Arroyo et al. used a CRISPR-based screen and consequently identified essential genes for oxidative phosphorylation.

8There is a major difference between cell-specific and organism-specific gene essentiality. That is, essential gene sets for human cells can be significantly different from those for human development. CRISPR technology, despite being powerful, cannot be used as a reverse genetics approach in humans for gene essentiality studies. Nevertheless, exome sequencing, another recent breakthrough, enables the identification of human essential genes in vivo.

9Exome sequencing is considerably less expensive than whole-genome sequencing, and most Mendelian diseases are caused by genetic variations in protein-coding regions (exomes). The Exome Aggregation Consortium (ExAC) reported the exome sequences of 60 706 individuals, and the genetic diversity represents an average of one variant of every 8 bases of the exomes. Thus these variations are analogous to a genome-wide mutagenesis screening conducted in nature, similar to a transposon mutagenesis screening performed in the lab. Strikingly, 3230 genes contain near-complete depletion of protein-truncating variants, representing candidate human organism-level essential genes. Therefore, the number of essential gene in humans in DEG 15 has increased by >10-fold, primarily due to the use of CRISPR and exome sequencing technology (Table 1).

10Tn-seq technology has been successfully used in identifying essential genes in a large number of bacteria, and it has also been used in archaea and even a eukaryote. In comparison to the single gene knockout method, Tn-seq is less time-consuming and labor-intensive, because of the parallel nature in mutagenesis and insertion site determination. The invention of the Tn-seq method can date back to a study in which Venter and coworkers performed Sanger sequencing to determine transposon insertion sites in 1999. In 2009, two technologies, high-density transposon-mediated mutagenesis and high-throughput sequencing, were mature, creating conditions that enabled Tn-seq to be invented. Many variants of Tn-seq were proposed, such as TraDIS, INSeq, HITS and Tn-seq Circle. Here, we refer to these methods collectively as Tn-seq since they all involve transposon mutagenesis and sequencing.

11Tn-seq has been widely used in identifying essential genes in bacteria. Figure 1A shows that since 2009, when DEG 5 was published, most bacterial essential genes have been determined by Tn-seq, and the proportion of essential genes that are determined by Tn-seq has been increasing ever since. This is not surprising given the powerfulness, ease of use, and the efficiency of Tn-seq in performing essentiality screening. Another advantage of Tn-seq is that it identifies not only essential protein-coding genes, but also non-coding genomic elements. For instance, by using Tn-seq, a large number of non-coding genomic elements have been determined in Acinetobacter baumannii, Brevundimonas subvibrioides, Escherichia coli O157:H7, Mycobacterium tuberculosis, Mycoplasma pneumonia, Salmonella entericaserovar Typhimurium, Sphingomonas wittichii and Streptococcus pneumonia.

12In addition, Tn-seq has been used to determine essential genes in species other than bacteria. The methanogenic archaeon Methanococcus maripaludis S2 is an obligate anaerobic prokaryote that lives in oxygen-free environments. Sarmiento et al. used the Tn-seq method and identified 526 essential genes required for growth in rich medium, representing the first genome-wide gene essentiality screening in archaea. The second essentiality screening in archaea was conducted in Sulfolobus islandicus, and some archaea specific essential genes were identified. Moreover, Tn-seq was also used in identifying essential genes in a eukaryote. Severe malaria is caused by the apicomplexan parasite Plasmodium falciparum, a unicellular protozoan parasite of humans, and 680 genes were identified as essential for optimal growth of this parasite. Because of the widespread use of Tn-seq, the number of prokaryotic essential genes in DEG 15 has more than doubled compared to that of DEG 10 (Figure 1A).

13To facilitate the use of DEG, we developed a set of analysis modules in the current release. Essential genes are preferentially situated in the leading strand, rather than the lagging strand, mainly because of the decreased mutagenesis pressure resulting from the head-on collisions of transcription and replication machineries in the leading strand. We obtained replication origins and determined leading vs. lagging strands using the DoriC database. Users can examine essential gene distributions between leading and lagging strands, and clicking the pie graph will display a list of genes in leading or lagging strands (Figure 2A).

14Sub-cellular localization and operon information were obtained from the PSORTb v3.0 tool and the DOOR database, respectively. Clicking a species name, e.g. Bacillus subtilis, will display sub-cellular localization distributions of essential genes, and detailed gene information can be further examined by clicking on a particular cell compartment (Figure 2B). Other information includes orthologous groups, EC number, KEGG pathway and GO, as determined by eggNOG-mapper. Users can analyze the GO distributions, and enriched GO terms powered by GOATOOLs, and enriched KEGG pathways, obtained using clusterProfiler package in R language. The analysis results, including strand bias distribution, sub-cellular distribution, and enrichment analysis of GO and KEGG pathways, are visualized with ECharts.

15To analyze human essential genes, we developed a tool by which users can compare and contrast the essential gene sets between experiments, generate Venn diagrams to visualize the comparison, and obtain unions and intersections for the two gene sets by clicking the corresponding graph (Figure 2C). Furthermore, DEG 15 continues to provide customizable BLAST tools that allow users to perform species- and experiment-specific searches for a single gene, a list of genes, annotated or un-annotated genomes.

16The identification of essential genes in both prokaryotes and eukaryotes has attracted significant attention over the past decade, largely because of the practical implications of these studies. Bacterial essential genes are attractive drug targets, as inhibiting these genes can suppress bacterial survival. Interest on essentiality screenings has also been boosted by synthetic biology, which aims to make an artificial self-sustainable living cell. The minimal gene set of a bacterium is considered a chassis for further addition of other parts with desirable traits. An increasing number of essentiality screens are being performed in a context-specific manner. For instance, essential genes for cancer cells can reveal cancer-specific cellular processes, which are targets for cancer drugs. Determination of essential genes of A. baumannii revealed genes required for its infection and survival in the lung. Moreover, another important direction is the prediction of gene essentiality using bioinformatic approaches, e.g., based on metabolic models. Therefore, because of the theoretical implications of the minimal-gene-set concept and its practical uses, it is expected that the essential gene identifications will continue to be further advanced.

17Reverse genetics will continue to be indispensable for pinpointing gene functions. It is expected that single-gene knockout projects for the model organisms, such as mice and Arabidopsis thaliana, will soon be completed. Multiple ways to manipulate gene expression are available, such as those based on TetR/Pip-OFF repressible promoter system and RNA interference. From the aspect of technology, this is a golden era for essential-gene research, because of the availability of Tn-seq and CRISPR/Cas9. The two technologies enable the gene essentiality screenings in a wide range of cell types and species under diverse conditions. Therefore, we anticipate that the increase in the number of essential genes for many cell types under various conditions will be accelerated in the future. Therefore, we will continue to update DEG with high-quality human-curated data in a timely manner to keep pace with this rapidly developing field.

18DEG is accessible from essentialgene.org or tubic.org/deg. All DEG data is freely available to download.