Skip to main content
Download PDF

Small open reading frames: a comparative genetics approach to validation

BMC Genomics volume 24, Article number: 226 (2023) Cite this article

Abstract

Open reading frames (ORFs) with fewer than 100 codons are generally not annotated in genomes, although bona fide genes of that size are known. Newer biochemical studies have suggested that thousands of small protein-coding ORFs (smORFs) may exist in the human genome, but the true number and the biological significance of the micropeptides they encode remain uncertain. Here, we used a comparative genomics approach to identify high-confidence smORFs that are likely protein-coding. We identified 3,326 high-confidence smORFs using constraint within human populations and evolutionary conservation as additional lines of evidence. Next, we validated that, as a group, our high-confidence smORFs are conserved at the amino-acid level rather than merely residing in highly conserved non-coding regions. Finally, we found that high-confidence smORFs are enriched among disease-associated variants from GWAS. Overall, our results highlight that smORF-encoded peptides likely have important functional roles in human disease.

Peer Review reports

Introduction

Open reading frames (ORFs) are stretches of DNA bounded by a stop codon, and represent an important genomic feature used to discover protein-coding genes. Traditionally, ORFs with less than 100 codons, also known as small ORFs (smORFs), were excluded from gene annotations. This arbitrary cutoff was imposed by the challenges of distinguishing protein-coding smORFs from random noise, both experimentally and computationally. However, recent advances in next-generation sequencing and proteomics have identified many smORF-encoded proteins, termed micropeptides [1,2,3].

smORFs reside in various regions of the genome, including 5’ and 3’ untranslated regions (UTRs), intergenic regions, and even in regions previously annotated as non-coding RNAs (ncRNAs) [1, 4]. Studies have also identified diverse biological functions for a few human micropeptides. For example, mitoregulin (MTLN), a 56-amino-acid (aa) micropeptide, is highly expressed in skeletal and cardiac muscle cells and is involved in regulating mitochondrial complex assembly [5]. Additionally, phospholamban (52 aa), sarcolipin (31 aa), and myoregulin (46 aa) are well-characterized regulators of the cardiac sarco-/endoplasmic reticulum calcium ATPase, SERCA [6]. Notably, mutations in the PLN gene are implicated in human dilated cardiomyopathy and heart failure, underscoring its crucial role in cardiac function [7, 8]. As a result, there is growing interest in annotating and functionally characterizing smORFs.

Recent studies combining RNA-sequencing (RNA-seq) and ribosome profiling (Ribo-seq) have identified and annotated thousands of putative smORFs across different human cell lines [1, 4]. While this approach is a highly significant advance over merely identifying ORFs in the human genome, the fraction of those thousands of putative smORFs that are truly protein-coding remains unclear and the functions of the micropeptides they encode remain unknown. To begin to address those uncertainties, functional genomic studies have been performed using mass spectrometry and CRISPR-based mutagenesis screens. These CRISPR-based mutagenesis screens have identified hundreds of smORFs in cell growth phenotypes [4, 9]. Nonetheless, there is still a need to develop methods to validate protein-coding smORFs, since new smORFs are constantly being identified in humans and other species. Therefore, an alternative strategy for identifying bona fide smORFs, which has not been attempted previously, is to use the power of comparative genetics, both across human populations and species. We hypothesized that smORFs that are truly protein-encoding should behave similarly to known genes and biologically validated smORFs with respect to human genetic variation and evolutionary conservation.

To examine the patterns of human genetic variation in putative smORFs, we annotated previously published smORF datasets with variants from 71,702 whole-genome sequencing samples from the Genome Aggregation Database (gnomAD) and defined constraint in smORFs using the robust missense observed/expected upper bound fraction (MOEUF) score [10]. Additionally, we characterized smORFs under evolutionary constraint using Genome Evolutionary Rate Profiling (GERP) scores [11]. Together these analyses allowed us to curate a list of high-confidence smORFs. Finally, we examined the role of smORFs in human disease and showed that disease-associated variants identified in genome-wide association studies (GWAS) are enriched in our set of high-confidence smORFs.

Results

Overview of smORF datasets

We analyzed datasets from two recent studies that performed Ribo-seq to annotate ORFs across various human cell lines (Fig. 1A). From each dataset, we first obtained a set of high quality smORFs that were not isoforms of known genes. Chen et al. identified 15,236 putative smORFs, and we proceeded with 8,080 smORFs that were not isoforms of known genes [4]. Similarly, Martinez et al. reported a gold standard set of 2,689 smORFs that were also not isoforms of known genes [1]. After obtaining putative smORFs, we first filtered them based on amino acid length (range 10–150 amino acids) and selected the longest isoform per smORF. While studies have demonstrated that smORFs that overlap protein-coding genes out-of-frame, are translated independently and can encode functional proteins [12,13,14,15], in our analysis we decided to exclude such ORFs as it would be challenging to distinguish those that represent protein-coding smORFs versus annotated genes translated out-of-frame. Specifically, we excluded smORFs overlapping annotated exons in the RefSeq database [16] as well as those overlapping the Peptide Atlas database [17]. In the set of smORFs that passed our initial filtering criteria (n = 5,789), we observed 515 smORFs with exact matches that were reported by both studies (Fig. 1B) and an additional 739 smORFs that had imperfect overlap. This minimal overlap may be due to differences in cell types, Ribo-Seq protocols, and ORF annotation tools between the studies. Given the minimal overlap between the two datasets, we considered the entire set of filtered smORFs in our downstream analysis to identify high-confidence smORFs.

Fig. 1
figure 1

Overview of smORF datasets. A Study workflow. B Venn diagram showing the overlap between the predictions from Chen et al. and Martinez et al. study in the filtered smORFs dataset. C Venn diagram showing the overlap between the predictions from Chen et al. and Martinez et al. in the high-confidence smORFs dataset. Abbreviations: MOEUF, missense observed/expected upper bound fraction; GERP, Genome Evolutionary Rate Profiling; GWAS, genome-wide association studies

Full size image

Selection of high-confidence smORFs

Given the limited functional evidence for the smORF predictions, we combined human genetic variation and evolutionary conservation metrics to identify high-confidence smORFs. First, we annotated the filtered list of putative smORFs with human genetic variants from the gnomAD v3 database (N = 71,702 genomes) and benchmarked these against a set of known smORFs with extensive biological validation [5, 6, 18,19,20,21,22,23,24,25,26,27,28,29,30,31] (n = 28, Table S1), a set of RefSeq genes with amino acid length < 150 and a selection of RefSeq genes [16] with varying probability of loss-of-function intolerant (pLI) scores. pLI is a metric to define genes that are intolerant to loss-of-function (LoF) variants, with scores ranging from 0 to 1 for least to most intolerant [32]. A high pLI score (≥ 0.9) indicates the gene is under strong selective constraint. Consistent with the range of pLI scores in the selected RefSeq genes, we observed a clear pattern in the distribution of non-synonymous/synonymous (N/S) and loss-of function/synonymous (LoF/S) ratios, whereby the median ratios were significantly lower in genes with high pLI scores compared to genes with low pLI scores, as expected (Fig. 2A and B, right facets, Kruskal–Wallis PN/S < 2.2 × 10–16 and PLoF/S < 2.2 × 10–16). Next, we compared N/S and LoF/S ratios among the different sets of known and putative smORFs. Here, we observed significant differences in the distribution of N/S and LoF/S ratios across the subsets of smORFs (Fig. 2A and B, left facets, Kruskal–Wallis PN/S = 3.23 x 10-5 and PLoF/S = 0.00197). Notably, the N/S ratios of the smORF subsets were comparable to RefSeq genes with low pLI scores and higher than RefSeq genes with moderate and high pLI scores (Fig. 2A).

Fig. 2
figure 2

Selection of high-confidence smORFs. For selected RefSeq genes with varying pLI scores and amino acid length < 150 (left facets) and smORFs (right facets), violin and boxplots showing A N/S ratios, B LoF/S ratios, C MOEUF scores, and D GERP scores. MOEUF and GERP thresholds used to filter putative smORFs are shown by dashed red line. Selected RefSeq genes are segregated by pLI scores ranging from low (n = 400), moderate (n = 400) and high (n = 400) scores, and genes with less than 150 amino acids (n = 400). smORF subsets includes known smORFs (n = 28), putative smORFs unique to Chen et al. (n = 4,030) and Martinez et al. (n = 1,244) dataset, putative smORFs with exact matches reported by both Chen et al. and Martinez et al. (n = 515), and smORFs in both datasets with imperfect overlap (n = 739). Box plots display the first quartile, median and third quartile. Abbreviations: N, nonsynonymous; S, synonymous; LoF, loss-of-function; pLI, probability of loss-of-function intolerant

Full size image

We observed that the number of LoF variants per smORF was usually < 10, which is too few to yield robust loss-of-function OEUF scores [10]. Therefore, to examine constraint within putative smORFs, we instead utilized the comparable missense variation metric, MOEUF, as this could be calculated based on a much larger number of single nucleotide variants (SNVs). MOEUF is the conservative estimate of the upper bound of the confidence interval for the observed-to-expected missense variants ratio. Low MOEUF scores indicate relatively higher intolerance to missense variation, and high MOEUF scores indicate a relatively higher tolerance to missense variation or limited statistical power to detect constraint [10]. While MOEUF examines the constraint of peptide sequences in the context of human populations, evolutionary conservation metrics, such as GERP, identify constrained DNA sequences across multiple species. GERP is based on a multiple mammalian species alignment, with positive scores indicating evolutionary constraint [11]. Therefore, we used MOEUF and GERP scores to provide additional evidence of the biological significance in smORFs.

As expected, the pattern of pLI scores in RefSeq genes was consistent with the degree of constraint based on MOEUF and GERP scores (Fig. 2C and D). Furthermore, known smORFs had MOEUF and GERP scores within the same range as RefSeq genes and were comparable to RefSeq genes with amino acid length < 150, in line with the previous biological functional characterization of these smORFs. Therefore, we utilized the range of MOEUF and GERP scores observed in these biologically validated smORFs and RefSeq genes to set thresholds to further filter predicted smORFs. We selected smORFs with MOEUF scores ≤ 1.5 and GERP scores ≥ -1 as this combination of filters captured almost the entire set of known smORFs (25/28), RefSeq genes with amino acid length < 150 (389/400) and RefSeq genes (1192/1200) while filtering out a considerable proportion of the putative smORFs (Fig. 2C and D, red dashed line). Post filtering, we were left with 3,326 smORFs (31% of all putative smORFs) as our high-confidence set (Table S1). In this high-confidence set, we observed 265 smORFs that were reported by both the Chen et al. and Martinez et al. (Fig. 1C).

High-confidence smORFs conserved at the protein-coding level

While our previous analyses allowed us to filter away putative smORFs of lower confidence, they did not address whether the sequences of the high-confidence smORFs were constrained due to the need to maintain the sequence for the encoded micropeptide per se versus merely residing in a highly constrained genomic region. To address this, we next asked whether the remaining set of filtered smORFs showed evidence of conservation at the amino acid level relative to the genome-wide background rate. We extracted the genomic coordinates for the five “incorrect” reading frames, two on the same strand and three on the opposite strand, for the set of high-confidence smORFs and a set of RefSeq genes. We then compared the N/S ratios and MOEUF scores of the five “incorrect” reading frames to the correct reading frame, providing five negative controls for every smORF and RefSeq gene. As expected, the correct reading frame for RefSeq genes showed significantly lower median N/S ratios (paired Wilcoxon signed rank test, all P values < 0.01) and MOEUF scores (paired Wilcoxon signed rank test, all P values < 0.02) compared to the other five reading frames. Similarly, we observed that high-confidence smORFs’ correct reading frame had significantly lower median N/S ratios and MOEUF scores than the “incorrect” reading frames (paired Wilcoxon signed rank test, all P < 0.02. The N/S ratios and MOEUF scores reflect stronger conservation in RefSeq genes (median GeneN/S = 2.23, GeneMOEUF = 0.822) compared to smORFs (median smORFN/S = 2.25, smORFMOEUF = 1.16), possibly driven by a fraction of smORFs within our high-confidence set that are not true protein-coding genes or by limited statistical power of MOEUF and N/S ratio to detect conserved smORFs (Fig. 3A and B).

Fig. 3
figure 3

Conservation of high-confidence smORFs at the protein-coding level. A Violin and box plots showing the distribution of N/S ratios of the correct (blue) and “incorrect” (grey) reading frames of RefSeq genes (n = 1200) and high-confidence smORFs (n = 2,891), with a lower N/S ratio observed in the correct reading. B Violin plot and boxplot showing that the correct reading frame (blue) had lower MOEUF scores compared to all five “incorrect” reading frames (grey) for both RefSeq genes (n = 1200) and high-confidence smORFs (n = 2,891). Box plots displaying the first quartile, median and third quartile. All P values are based on the paired Wilcoxon signed rank test

Full size image

Enrichment of disease-associated GWAS variants in high-confidence smORFs

We next explored the possible role of smORFs in human disease. We investigated whether common coding variants (MAF ≥ 0.01) in high-confidence smORFs showed evidence of association with human traits via a GWAS enrichment analysis. To do this, we identified common SNVs reported in the GWAS catalog that fell within coding regions of high-confidence smORFs [33]. We identified 25 SNVs overlapping smORFs with genome-wide significance (P < 5 × 10–8), including 18 non-synonymous and seven synonymous SNVs. These variants are implicated in a wide range of traits such as HDL cholesterol, HIV infection, breast cancer, and tuberculosis (Table S2). Next, we assessed whether the number of GWAS SNVs overlapping smORFs was greater than expected by chance. We compiled SNV lists from four SNV arrays commonly used in GWAS and identified all SNVs within ± 1 Mb that were in high linkage disequilibrium (r2 ≥ 0.8) of the genotyped variants on those four arrays [34, 35]. We then separated the SNVs into bins by minor allele frequency (MAF) and ran five separate permutation tests, one per MAF bin. Through the series of permutation tests, we observed a significant enrichment of trait-associated SNVs within our high-confidence set of smORFs (Fig. 4, P = 4.96 × 10–5, 10,000 permutations). We next explored the functional impact of the GWAS SNVs on the putative smORFs by computing Polyphen-2 scores ([36]. Similar to prior work in annotated genes [37], we identified no significant enrichment of nonsynonymous-damaging SNVs after performing a permutation analysis (Ndamaging = 9, Ntotal = 18, Fisher’s exact test P = 0.36).

Fig. 4
figure 4

Enrichment of GWAS variants within smORFs. Permutation analysis testing enrichment of SNVs associated with disease/ other traits within smORFs segregated by MAF bins (facets) revealed a statistically significant enrichment among smORFs (Fisher’s method meta-analysis P = 4.96 × 10–5). P values were calculated by comparing the number of observed overlapping SNVs (red) to the number expected based on 10,000 permutations (grey histograms). Meta-analysis using Fisher’s method, H0: no significant enrichment of GWAS SNVs in smORFs. Abbreviation: MAF, minor allele frequency

Full size image

Discussion

In this study, we utilized a comparative genomics approach to identify a set of high-confidence smORFs that are likely encoding functional micropeptides. We used datasets from two previously published studies that performed Ribo-seq to annotate novel smORFs [1, 4]. However, what remained unclear is the extent to which their predictions represented proteins with independent functions since ORFs can be translated as part of regulatory mechanisms or can be translational noise. Here, we provide further evidence of the functional significance of the predicted smORFs through analysis of constraint within human populations and evolutionary conservation. We compared the MOEUF and GERP score distributions of the predicted smORFs to annotated RefSeq genes, and after imposing strict thresholds, we generated a list of 3,326 high-confidence smORFs. High confidence smORFs tend to have similar distributions to less conserved genes, i.e., those with low and moderate pLI scores. We speculate this is because (1) smORFs may have a younger evolutionary age (younger genes tend to be shorter) and (2) lower statistical power to detect highly conserved genes [38,39,40,41,42]. Additionally, we assessed conservation of these high-confidence smORFs relative to the genome-wide background rate. By comparing the relative frequency of functional versus non-functional protein-coding variation, we showed that our high-confidence set is enriched with smORFs conserved at the protein-coding level, rather than merely being highly conserved non-coding regions. These results indicate that selective pressures are acting on the amino acid sequence of this set of high-confidence smORFs.

To explore the biological significance of our high-confidence smORFs, we asked whether genetic variation in these smORFs contributes to human traits and diseases. We showed that disease-associated SNVs with genome-wide significance are enriched in our high-confidence smORFs. In addition, we explored the functional impact of the GWAS SNVs on the smORFs by computing Polyphen-2 scores. Similar to prior work in annotated genes, we identified no significant enrichment of nonsynonymous-damaging variants among those implicated in human disease [37]. The reason for the lack of significant enrichment of damaging SNVs could be because (1) the GWAS SNVs used in our analysis are not the causal variants but rather are in linkage disequilibrium with causal SNVs, (2) Polyphen-2 predicted that the majority of smORF variants were damaging, limiting our statistical power, or (3) Polyphen-2 was developed for larger genes and might not be well-calibrated for smORFs. In this study, we analyze common variants (MAF ≥ 0.01), which have been the historical focus of GWAS. In the future, similar analyses can be applied to potential association results from large-scale whole genome sequencing efforts, which may include more rare variants [43, 44].

Finally, given that smORFs are commonly found in 5’ UTRs, 3’ UTRs, and introns of annotated genes, future work would need to leverage expression and proteomic molecular data to test if the GWAS associations with smORFs are mediated by SNVs that directly affect the smORF and/or nearby genes. A common follow-up analysis used to functionally evaluate GWAS signals is to perform statistical co-localization analysis. This analysis combines GWAS findings with functional data, such as expression quantitative trait loci, to examine if the same variant is causal in both [45, 46]. Currently, smORFs are not well-represented in most tissue and cell-specific expression datasets, such as the Genotype-Tissue Expression database [47], limiting the follow-up analyses we could perform to functionally interpret our GWAS findings.

To conclude, we present a robust workflow to identify high-confidence human smORFs using a comparative genetics approach. The alternate reading-frame analysis allowed us to assess conservation of the set of high-confidence smORFs at the protein-coding level and was critical to differentiating protein-coding regions from highly conserved non-coding regions. In addition, we found that smORFs are significantly enriched with disease-associated GWAS SNVs. Future studies analyzing novel human smORFs can follow our workflow (Fig. 1a), utilizing the empirical cutoffs for MOEUF and GERP scores as these are based on comparisons to known RefSeq genes. Since, our workflow leverages large population databases, we would only recommend our approach for analysis in other organisms where large whole-genome sequencing databases are available. Overall, our results are a step forward in showing that smORFs likely encode functional proteins implicated in human disease. Further advancements in experimental and computational tools are needed to improve annotation of novel smORFs and uncover their role in biology [48].

Methods

Dataset filtering

We downloaded datasets of predicted protein-coding ORFs from two previous published studies [1, 4], performed liftover to convert the datasets from hg19 to hg38 using the default parameters of UCSC Genome Browser’s web-interface liftOver tool, removed ORFs that were less than 10 codons and greater than 150 codons, and filtered to keep the longest coding DNA sequence for each predicted ORF [49]. For the Chen et al. dataset, we only considered smORFs that were labelled as upstream, new, start overlap, stop overlap, new isoform, downstream, and long out of frame [4]. Next, we intersected smORF exon regions with annotated RefSeq genes and the peptide atlas 'human Jan 2020' database, a database of peptides identified from tandem mass spectrometry experiments, using bedtools intersect [16, 17, 50]. We removed smORFs where the exon regions had a ≥ 50% overlap with exons of annotated RefSeq genes or vice versa. Additionally, smORFs where the exons regions had a 100% overlap with genomic coordinates of peptides from the Peptide Atlas database and ≥ 50% overlap with RefSeq genes were excluded. In addition to the initial filtering, we intersected the datasets to examine the similarity between the predictions from both studies. We performed two intersections, one to identify smORF predictions from both studies that were identical and the other considering a minimum of 1-basepair overlap.

Constraint calculation

We adopted previous workflows to calculate MOEUF scores for known genes and novel smORFs [10]. We downloaded the gnomAD Hail table of all possible genomic variation with methylation and CpG context, filtered for regions in smORFs and RefSeq Select genes, performed liftover of Hail tables from hg19 to hg38, used ANNOVAR to annotate the "all possible variation" table with smORFs/gene function and gnomAD v3 WGS allele frequencies (based on 71,702 samples), identified all observed variants with minor allele frequency (MAF) < 0.001, used the mutation rate table to calculate expected variants based on 3-mer sequence context and finally calculated MOEUF scores for every smORF and RefSeq Select gene [16, 50,51,52]. Differences between this workflow and previously published pipelines to calculate MOEUF include using WGS data without coverage correction instead of whole-exome sequencing, ANNOVAR instead of Variant Effect Predictor, and obtaining the OEUF standard deviation from the combined set of all genes and smORFs. We validated this pipeline by comparing expected and observed variants for known genes to previously published scores and by calculating the correlation between expected and observed synonymous variants. We observed a high correlation between the number of expected missense variants per gene (r = 0.98) among 15,010 genes in both our results and the published score database. Consistent with prior correlations between the number of expected and observed synonymous variants per gene, we observed a high correlation (r = 0.98) across 30,334 genes and smORFs.

Evolutionary conservation analysis

We extracted base-wise GERP scores for the list of smORFs from the GERP track on the UCSC Genome Table Browser [11, 53]. The average GERP score across smORF exons was calculated using bedtools intersect and bedtools group by [50].

Damaging variant score

Polyphen-2 predicts the structural and functional impact of a missense substitution and assigns a score indicating the probability of the substitution being damaging. Scores range from 0 (benign) to 1 (damaging). Scores > 0.85 are more confidently predicted to be damaging and are termed ‘probably damaging’. smORF genetic variants were first functionally annotated using ANNOVAR’s gene-based annotation [52]. To compute the scores, the annotated variants were batch queried on a custom Polyphen-2 server [36].

GWAS analysis and permutation testing

We filtered the NHGRI-EBI GWAS catalog, removing redundant single nucleotide polymorphisms (SNVs) and retaining only biallelic SNVs with a p-value < 5 × 10–8 (n = 57,099) [33, 52]. We downloaded all observed variants from gnomAD v3 database for every smORF and functionally annotated the variants using ANNOVAR [51, 52]. We then searched the annotated smORF variants against the GWAS catalog to identify SNVs associated with disease/traits. Next, we performed a permutation test to assess whether the enrichment of SNVs identified by GWAS in smORFs was significant. We compiled SNV lists from four SNV arrays (Affymetrix SNV 6.0, Illumina Human1M-Duo, Illumina HumanOmni1_Quad, and Illumina HumanHap550), accessed through the SNV/CNV track on UCSC Genome Table Browser [53], performed liftover to convert these from hg19 to hg38, removed redundant SNVs and filtered to keep biallelic SNVs (n = 1,569,244). Using plink1.9 software, we identified all SNVs that were in linkage disequilibrium (r2 ≥ 0.8) and within 1 Mb of the SNVs in the SNV lists [34, 35]. We then separated the compiled SNV list into bins by MAF (1–10%, 10–20%, 20–30%, 30–40% and 40–50%) and ran five separate permutation tests, one per MAF bin. We sampled variants with replacement from each SNV list, corresponding to the number of variants from the GWAS catalog in each MAF bin, and calculated the number of SNVs overlapping smORF exons. For each test, we computed the p-value based on 10,000 permutations and the observed number of SNVs identified by GWAS overlapping smORFs. We computed the meta-analysis p-value using Fisher's method, which combines p-values from independent tests into a single test statistic.

In addition, we performed a permutation analysis testing whether there was an enrichment of SNVs scored as damaging by Polyphen-2 amongst the GWAS SNV hits overlapping our high-confidence smORFs. We sampled variants with replacement from a set of Polyphen-2 scored nonsynonymous smORF SNVs (MAF 1–50%) and computed the p-value based on 10,000 permutations and the observed number of GWAS SNVs scored as damaging.

Availability of data and materials

The Martinez et al. [1] dataset analyzed in this study is included in the supplementary information (supplementary data 1) of this published article, Martinez TF, Chu Q, Donaldson C, Tan D, Shokhirev MN, Saghatelian A: Accurate annotation of human protein-coding small open reading frames. Nat Chem Biol 2020, 16(4):458–468.

The Chen et al. [4] dataset analyzed in this study is available in the GEO database, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE131650 (GSE131650_orfratings.bed.gz).

Filtered datasets are within the manuscript and its supporting information files.

References

  1. Martinez TF, Chu Q, Donaldson C, Tan D, Shokhirev MN, Saghatelian A. Accurate annotation of human protein-coding small open reading frames. Nat Chem Biol. 2020;16(4):458–68.

    Article  CAS  PubMed  Google Scholar 

  2. Couso JP. Finding smORFs: getting closer. Genome Biol. 2015;16:189.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Basrai MA, Hieter P, Boeke JD. Small open reading frames: beautiful needles in the haystack. Genome Res. 1997;7(8):768–71.

    Article  CAS  PubMed  Google Scholar 

  4. Chen J, Brunner AD, Cogan JZ, Nunez JK, Fields AP, Adamson B, Itzhak DN, Li JY, Mann M, Leonetti MD, et al. Pervasive functional translation of noncanonical human open reading frames. Science. 2020;367(6482):1140–6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Stein CS, Jadiya P, Zhang X, McLendon JM, Abouassaly GM, Witmer NH, Anderson EJ, Elrod JW, Boudreau RL. Mitoregulin: A lncRNA-Encoded Microprotein that Supports Mitochondrial Supercomplexes and Respiratory Efficiency. Cell Rep. 2018;23(13):3710-3720e3718.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Anderson DM, Makarewich CA, Anderson KM, Shelton JM, Bezprozvannaya S, Bassel-Duby R, Olson EN. Widespread control of calcium signaling by a family of SERCA-inhibiting micropeptides. Sci Signal. 2016;9(457):ra119.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Eijgenraam TR, Boukens BJ, Boogerd CJ, Schouten EM, van de Kolk CWA, Stege NM, Te Rijdt WP, Hoorntje ET, van der Zwaag PA, van Rooij E, et al. The phospholamban p.(Arg14del) pathogenic variant leads to cardiomyopathy with heart failure and is unreponsive to standard heart failure therapy. Sci Rep. 2020;10(1):9819.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Schmitt JP, Kamisago M, Asahi M, Li GH, Ahmad F, Mende U, Kranias EG, MacLennan DH, Seidman JG, Seidman CE. Dilated cardiomyopathy and heart failure caused by a mutation in phospholamban. Science. 2003;299(5611):1410–3.

    Article  CAS  PubMed  Google Scholar 

  9. Prensner JR, Enache OM, Luria V, Krug K, Clauser KR, Dempster JM, et al. Noncanonical open reading frames encode functional proteins essential for cancer cell survival. Nat Biotechnol. 2021;39(6):697–704.

  10. Karczewski KJ, Francioli LC, Tiao G, Cummings BB, Alfoldi J, Wang Q, Collins RL, Laricchia KM, Ganna A, Birnbaum DP, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581(7809):434–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Cooper GM, Stone EA, Asimenos G, Program NCS, Green ED, Batzoglou S, Sidow A. Distribution and intensity of constraint in mammalian genomic sequence. Genome Res. 2005;15(7):901–13.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Cao X, Khitun A, Luo Y, Na Z, Phoodokmai T, Sappakhaw K, Olatunji E, Uttamapinant C, Slavoff SA. Alt-RPL36 downregulates the PI3K-AKT-mTOR signaling pathway by interacting with TMEM24. Nat Commun. 2021;12(1):508.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Cao X, Khitun A, Harold CM, Bryant CJ, Zheng SJ, Baserga SJ, Slavoff SA. Nascent alt-protein chemoproteomics reveals a pre-60S assembly checkpoint inhibitor. Nat Chem Biol. 2022;18(6):643–51.

    Article  CAS  PubMed  Google Scholar 

  14. Gagnon M, Savard M, Jacques JF, Bkaily G, Geha S, Roucou X, Gobeil F. Potentiation of B2 receptor signaling by AltB2R, a newly identified alternative protein encoded in the human bradykinin B2 receptor gene. J Biol Chem. 2021;296: 100329.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Brunet MA, Jacques JF, Nassari S, Tyzack GE, McGoldrick P, Zinman L, Jean S, Robertson J, Patani R, Roucou X. The FUS gene is dual-coding with both proteins contributing to FUS-mediated toxicity. EMBO Rep. 2021;22(1): e50640.

    Article  CAS  PubMed  Google Scholar 

  16. O’Leary NA, Wright MW, Brister JR, Ciufo S, Haddad D, McVeigh R, Rajput B, Robbertse B, Smith-White B, Ako-Adjei D, et al. Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res. 2016;44(D1):D733-745.

    Article  CAS  PubMed  Google Scholar 

  17. Deutsch EW. The PeptideAtlas Project. Methods Mol Biol. 2010;604:285–96.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Makarewich CA, Baskin KK, Munir AZ, Bezprozvannaya S, Sharma G, Khemtong C, Shah AM, McAnally JR, Malloy CR, Szweda LI, et al. MOXI Is a mitochondrial micropeptide that enhances fatty acid beta-oxidation. Cell Rep. 2018;23(13):3701–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Bi P, McAnally JR, Shelton JM, Sanchez-Ortiz E, Bassel-Duby R, Olson EN. Fusogenic micropeptide Myomixer is essential for satellite cell fusion and muscle regeneration. Proc Natl Acad Sci U S A. 2018;115(15):3864–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Quinn ME, Goh Q, Kurosaka M, Gamage DG, Petrany MJ, Prasad V, Millay DP. Myomerger induces fusion of non-fusogenic cells and is required for skeletal muscle development. Nat Commun. 2017;8:15665.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Zhang Q, Vashisht AA, O’Rourke J, Corbel SY, Moran R, Romero A, Miraglia L, Zhang J, Durrant E, Schmedt C, et al. The microprotein Minion controls cell fusion and muscle formation. Nat Commun. 2017;8:15664.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Bi P, Ramirez-Martinez A, Li H, Cannavino J, McAnally JR, Shelton JM, Sanchez-Ortiz E, Bassel-Duby R, Olson EN. Control of muscle formation by the fusogenic micropeptide myomixer. Science. 2017;356(6335):323–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Matsumoto A, Pasut A, Matsumoto M, Yamashita R, Fung J, Monteleone E, Saghatelian A, Nakayama KI, Clohessy JG, Pandolfi PP. mTORC1 and muscle regeneration are regulated by the LINC00961-encoded SPAR polypeptide. Nature. 2017;541(7636):228–32.

    Article  CAS  PubMed  Google Scholar 

  24. D’Lima NG, Ma J, Winkler L, Chu Q, Loh KH, Corpuz EO, Budnik BA, Lykke-Andersen J, Saghatelian A, Slavoff SA. A human microprotein that interacts with the mRNA decapping complex. Nat Chem Biol. 2017;13(2):174–80.

    Article  CAS  PubMed  Google Scholar 

  25. Anderson DM, Anderson KM, Chang CL, Makarewich CA, Nelson BR, McAnally JR, Kasaragod P, Shelton JM, Liou J, Bassel-Duby R, et al. A micropeptide encoded by a putative long noncoding RNA regulates muscle performance. Cell. 2015;160(4):595–606.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Pauli A, Norris ML, Valen E, Chew GL, Gagnon JA, Zimmerman S, Mitchell A, Ma J, Dubrulle J, Reyon D, et al. Toddler: an embryonic signal that promotes cell movement via Apelin receptors. Science. 2014;343(6172):1248636.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Chng SC, Ho L, Tian J, Reversade B. ELABELA: a hormone essential for heart development signals via the apelin receptor. Dev Cell. 2013;27(6):672–80.

    Article  CAS  PubMed  Google Scholar 

  28. Magny EG, Pueyo JI, Pearl FM, Cespedes MA, Niven JE, Bishop SA, Couso JP. Conserved regulation of cardiac calcium uptake by peptides encoded in small open reading frames. Science. 2013;341(6150):1116–20.

    Article  CAS  PubMed  Google Scholar 

  29. Shanmugam M, Molina CE, Gao S, Severac-Bastide R, Fischmeister R, Babu GJ. Decreased sarcolipin protein expression and enhanced sarco(endo)plasmic reticulum Ca2+ uptake in human atrial fibrillation. Biochem Biophys Res Commun. 2011;410(1):97–101.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. MacLennan DH, Kranias EG. Phospholamban: a crucial regulator of cardiac contractility. Nat Rev Mol Cell Biol. 2003;4(7):566–77.

    Article  CAS  PubMed  Google Scholar 

  31. Wright BW, Yi Z, Weissman JS, Chen J. The dark proteome: translation from noncanonical open reading frames. Trends Cell Biol. 2022;32(3):243–58.

    Article  CAS  PubMed  Google Scholar 

  32. Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, O’Donnell-Luria AH, Ware JS, Hill AJ, Cummings BB, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536(7616):285–91.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, McMahon A, Morales J, Mountjoy E, Sollis E, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005–12.

    Article  CAS  PubMed  Google Scholar 

  34. Renteria ME, Cortes A, Medland SE. Using PLINK for Genome-Wide Association Studies (GWAS) and data analysis. Methods Mol Biol. 2013;1019:193–213.

    Article  PubMed  Google Scholar 

  35. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81(3):559–75.

  36. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS, Sunyaev SR. A method and server for predicting damaging missense mutations. Nat Methods. 2010;7(4):248–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Chen R, Davydov EV, Sirota M, Butte AJ. Non-synonymous and synonymous coding SNPs show similar likelihood and effect size of human disease association. PLoS ONE. 2010;5(10): e13574.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Carvunis AR, Rolland T, Wapinski I, Calderwood MA, Yildirim MA, Simonis N, Charloteaux B, Hidalgo CA, Barbette J, Santhanam B, et al. Proto-genes and de novo gene birth. Nature. 2012;487(7407):370–4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Wolf YI, Novichkov PS, Karev GP, Koonin EV, Lipman DJ. The universal distribution of evolutionary rates of genes and distinct characteristics of eukaryotic genes of different apparent ages. Proc Natl Acad Sci U S A. 2009;106(18):7273–80.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Zheng EB, Zhao L: Protein evidence of unannotated ORFs in drosophila reveals diversity in the evolution and properties of young proteins. Elife. 2022;11:e78772.

  41. Vakirlis N, Vance Z, Duggan KM, McLysaght A. De novo birth of functional microproteins in the human lineage. Cell Rep. 2022;41(12): 111808.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Duffy EE, Finander B, Choi G, Carter AC, Pritisanac I, Alam A, Luria V, Karger A, Phu W, Sherman MA, et al. Developmental dynamics of RNA translation in the human brain. Nat Neurosci. 2022;25(10):1353–65.

    Article  CAS  PubMed  Google Scholar 

  43. Ramirez AH, Sulieman L, Schlueter DJ, Halvorson A, Qian J, Ratsimbazafy F, Loperena R, Mayo K, Basford M, Deflaux N, et al. The All of Us Research Program: Data quality, utility, and diversity. Patterns (N Y). 2022;3(8): 100570.

    Article  PubMed  Google Scholar 

  44. Taliun D, Harris DN, Kessler MD, Carlson J, Szpiech ZA, Torres R, Taliun SAG, Corvelo A, Gogarten SM, Kang HM, et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program. Nature. 2021;590(7845):290–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. de Goede OM, Nachun DC, Ferraro NM, Gloudemans MJ, Rao AS, Smail C, Eulalio TY, Aguet F, Ng B, Xu J, et al. Population-scale tissue transcriptomics maps long non-coding RNAs to complex disease. Cell. 2021;184(10):2633-2648 e2619.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Wu Y, Broadaway KA, Raulerson CK, Scott LJ, Pan C, Ko A, He A, Tilford C, Fuchsberger C, Locke AE, et al. Colocalization of GWAS and eQTL signals at loci with multiple signals identifies additional candidate genes for body fat distribution. Hum Mol Genet. 2019;28(24):4161–72.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Carithers LJ, Moore HM. The Genotype-Tissue Expression (GTEx) Project. Biopreserv Biobank. 2015;13(5):307–8.

    Article  PubMed  PubMed Central  Google Scholar 

  48. Mudge JM, Ruiz-Orera J, Prensner JR, Brunet MA, Calvet F, Jungreis I, Gonzalez JM, Magrane M, Martinez TF, Schulz JF, et al. Standardized annotation of translated open reading frames. Nat Biotechnol. 2022;40(7):994–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Kuhn RM, Haussler D, Kent WJ. The UCSC genome browser and associated tools. Brief Bioinform. 2013;14(2):144–61.

    Article  CAS  PubMed  Google Scholar 

  50. Quinlan AR, Hall IM. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics. 2010;26(6):841–2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Karczewski KJ, Francioli LC, Tiao G, Cummings BB, Alfoldi J, Wang Q, et al. Author correction: the mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2021;590(7846):E53.

  52. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16): e164.

    Article  PubMed  PubMed Central  Google Scholar 

  53. Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, Kent WJ. The UCSC Table Browser data retrieval tool. Nucleic Acids Res. 2004;32(Database issue):D493-496.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank Dr. Gabriel Hoffman for his valuable input regarding the analysis validating that smORFs are conserved at the protein-coding level. Research reported in this paper was supported by the Office of Research Infrastructure of the National Institutes of Health under award number S10OD018522. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was supported in part through the computational resources and staff expertise provided by Scientific Computing at the Icahn School of Medicine at Mount Sinai. This paper is based on a thesis submitted in fulfillment of the requirements for the MS degree in Biomedical Sciences at the Graduate School of Biomedical Sciences, Icahn School of Medicine at Mount Sinai.

Funding

NA.

Author information

Author notes

  1. Niyati Jain and Felix Richter contributed equally as first author.

  2. Andrew J. Sharp and Bruce D. Gelb contributed equally as senior author.

Authors and Affiliations

  1. Department of Genetics and Genomic Sciences and Mindich Child Health and Development Institute, Icahn School of Medicine at Mount, Hess Center for Science and Medicine, 1470 Madison Avenue, New York, NY, 10029, USA

    Niyati Jain, Andrew J. Sharp & Bruce D. Gelb

  2. Present Address: Committee On Genetics, Genomics, and Systems Biology, The University of Chicago, Chicago, IL, USA

    Niyati Jain

  3. Department of Pediatrics, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    Felix Richter & Bruce D. Gelb

  4. Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA

    Ivan Adzhubei

  5. Division of Genetics, Brigham and Women’s Hospital, Boston, MA, USA

    Ivan Adzhubei

Authors
  1. Niyati Jain
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Felix Richter
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Ivan Adzhubei
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Andrew J. Sharp
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Bruce D. Gelb
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

N.J. performed dataset filtering, computed GERP scores, and ran the GWAS enrichment analysis. F.R. advised N.J. on the technical aspects of running the different analyses. F.R. performed the constraint calculation (computing MOEUF scores). N.J. prepared Figs. 1, 23 and 4. I.A. built a custom Polyphen-2 server and assisted in computing the scores for the high-confidence smORFs. N.J. and F.R. interpreted the data and wrote the main manuscript text. A.J.S. and B.D.G. were major contributors to the conception of the work, interpretation of the data, and editing of the main manuscript text. All authors reviewed and approved the manuscript.

Corresponding author

Correspondence to Bruce D. Gelb.

Ethics declarations

Ethics approval and consent to participate

No administrative permissions were required to access the raw data used in this study.

Consent for publication

NA.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1:

 Table S1. List of high-confidence smORFs and known smORFs.

Additional file 2:

 Table S2. List of smORFs carrying disease-associated SNVs from GWAS catalog.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jain, N., Richter, F., Adzhubei, I. et al. Small open reading frames: a comparative genetics approach to validation. BMC Genomics 24, 226 (2023). https://doi.org/10.1186/s12864-023-09311-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12864-023-09311-7

Keywords

BMC Genomics

ISSN: 1471-2164

Contact us