Features of Functional Human Genes

Abstract

Proteins and non-coding RNAs are functional products of the genome that carry out the bulk of crucial cellular processes. With recent technological advances, researchers can sequence genomes in the thousands as well as probe for specific genomic activities of multiple species and conditions. These studies have identified thousands of potential proteins, RNAs and associated activities, however there are conflicting conclusions on the functional implications depending upon the burden of evidence researchers use, leading to diverse interpretations of which regions of the genome are "functional". Here we investigate the association between gene functionality and genomic features, by comparing established functional protein-coding and non-coding genes to non-genic regions of the genome. We find that the strongest and most consistent association between functional genes and any genomic feature is evolutionary conservation and transcriptional activity. Other strongly associated features include sequence alignment statistics, such as maximum between-site covariation. We have also identified some concerns with 1,000 Genomes Project and Genome Aggregation Database SNP densities, as short non-coding RNAs tend to have greater than expected SNP densities. Our results demonstrate the importance of evolutionary conservation and transcription for sequence functionality, which should both be taken into consideration when differentiating between functional sequences and noise.

Spotlight on alternative frame coding: Two long overlapping genes in Pseudomonas aeruginosa are translated and under purifying selection

“…Multiple independent lines of evidence indicated that both OLGs are under purifying (negative) selection, a strong indicator of functionality, particularly when combined with evidence of expression ( Cooper and Gardner, 2020 ). Firstly, the ORFs for olg1 and olg2 were both significantly longer than expected, given the amino acids (AA) in the reference genes tle3 and PA1383, using the synonymous-mutation method from the tool “Frameshift” ( Schlub et al., 2018 ) ( Figure 5 B).…”

Flawed machine-learning confounds coding sequence annotation

“…We have highlighted the scope for further development of tools to address the coding sequence annotation challenge. We note that the existing tools have not employed some of the latest machine-learning methods, nor do they include more complex information for conserved peptide predictions, for example the increased covariation in coding sequence alignments has been observed in recent analyses [50, 51]. We also note there has not been much work in identifying the strengths of the statistical features that contribute to predictions.…”

Thirteen dubious ways to detect conserved structural RNAs

“…That is in contrast with RNA structural covariations which are predominantly distant (95% of covariation in S. cerevisiae structural ncRNAs are more than three nucleotides apart) 91 . Other studies have also reported coding covariations 92 . Significant covariation between adjacent or next to adjacent positions is a signal indicative of protein‐coding exons, not yet exploited, that could be incorporated into methods to detect protein‐coding exons from alignments.…”

“…91 Other studies have also reported coding covariations. 92 Significant covariation between adjacent or next to adjacent positions is a signal indicative of protein-coding exons, not yet exploited, that could be incorporated into methods to detect protein-coding exons from alignments.…”