Identification and characterisation of endogenous Avian Leukosis Virus subgroup E (ALVE) insertions in chicken whole genome sequencing data

Abstract

Background: Endogenous retroviruses (ERVs) are the remnants of retroviral infections which can elicit prolonged genomic and immunological stress on their host organism. In chickens, endogenous Avian Leukosis Virus subgroup E (ALVE) expression has been associated with reductions in muscle growth rate and egg production, as well as providing the potential for novel recombinant viruses. However, ALVEs can remain in commercial stock due to their incomplete identification and association with desirable traits, such as ALVE21 and slow feathering. The availability of whole genome sequencing (WGS) data facilitates high-throughput identification and characterisation of these retroviral remnants. Results: We have developed obsERVer, a new bioinformatic ERV identification pipeline which can identify ALVEs in WGS data without further sequencing. With this pipeline, 20 ALVEs were identified across eight elite layer lines from Hy-Line International, including four novel integrations and characterisation of a fast feathered phenotypic revertant that still contained ALVE21. These bioinformatically detected sites were subsequently validated using new highthroughput KASP assays, which showed that obsERVer was highly precise and exhibited a 0% false discovery rate. A further fifty-seven diverse chicken WGS datasets were analysed for their ALVE content, identifying a total of 322 integration sites, over 80% of which were novel. Like exogenous ALV, ALVEs show site preference for proximity to protein-coding genes, but also exhibit signs of selection against deleterious integrations within genes. Conclusions: obsERVer is a highly precise and broadly applicable pipeline for identifying retroviral integrations in WGS data. ALVE identification in commercial layers has aided development of high-throughput diagnostic assays which will aid ALVE management, with the aim to eventually eradicate ALVEs from high performance lines. Analysis of non-commercial chicken datasets with obsERVer has revealed broad ALVE diversity and facilitates the study of the biological effects of these ERVs in wild and domesticated populations.

Falling fowl of the chicken reference genome: pitfalls of studying polymorphic endogenous retroviruses

“…Conversely, the highly expressed ALVE6 (indicating its order of discovery in White Leghorn chickens in the 1980s) is truncated to just the envelope and 3′LTR, and is widespread, yet polymorphic, among commercial layers and broilers, but has not been identified in other red junglefowl. Excluding the low numbers observed in highly-selected Leghorns, ALVE abundance is typically in excess of six integrations per genome, and is usually > 10 in non-commercial lines and available red junglefowl datasets [ 12 , 13 ]. Previous work has shown that despite the morphological and behavioural characteristics of the reference individual [ 14 ], this genome is heavily introgressed with the White Leghorn breed and not representative of wild red junglefowl, modern or ancestral [ 15 ].…”

“…However, the authors attribute all ALVE expression specifically to ALVE1, without confirming its presence in the genome. In fairness, all birds and cell lines used in this study were derived from White Leghorns, where ALVE1 is highly prevalent yet still polymorphic, even within individual flocks [ 12 ]. Furthermore, most White Leghorns contain 3 or more ALVE integrations, and the common ALVE3, ALVE6 and ALVE9 elements all exhibit high envelope expression.…”

“…Fortunately, as high-throughput sequencing approaches have become more accurate and cost-effective, this has become more achievable. ALVE integrations can be detected confidently from whole genome sequencing data [ 12 ], utilising enrichment approaches to exclusively assess study population ALVE diversity if budgets require [ 19 ]. Furthermore, these approaches are broadly applicable to ERVs across vertebrate genomes.…”

Le projet « Mille Génomes Gallus » : partager les données de séquences pour mieux les utiliser

“…-l'Université de Wageningen aux Pays-Bas (WU) et de l'entreprise Hendrix Genetics, qui disposent déjà d'un total d'au moins 250 séquences individuelles ayant une profondeur au moins égale à 10X ; l'entreprise Hendrix Genetics a financé l'acquisition de la plus grande partie des données proposées et s'est montrée ouverte à l'approche 1 000 génomes pour améliorer la connaissance du génome et de la diversité génétique, mais elle ne souhaitait pas que les données soient mutualisées pour la recherche de mutations causales déterminant les QTL ; -le Friedrich Loeffler Institute (FLI) avec son laboratoire de Mariensee, et l'Université de Göttingen (UGOE) en Allemagne, avec 50 séquences individuelles de profondeur égale ou supérieure à 10X ; cette équipe a déjà publié une analyse d'un jeu de 127 séquences obtenues sur des lignées commerciales et l'espèce sauvage Gallus gallus gallus, avec le principal objectif d'identifier des signatures de sélection ou de domestication par la comparaison des fréquences alléliques entre groupes contrastés (Qanbari et al, 2019). Les 2 régions qui apparaissent le plus soumises à sélection sont centrées soit autour du gène ALX1, impliqué dans la morphologie du bec, soit autour du gène KITLG, impliqué dans la pigmentation dans de nombreuses espèces mais pas encore chez la poule.…”

“…Le Roslin Institute détient une centaine de données de séquence et a manifesté un intérêt de principe. Récemment, ce groupe a publié une analyse de la variation des insertions rétrovirales endogènes à partir de 65 données de séquence de génome entier, la plupart provenant de séquences réalisées en mélanges d'ADN (39 mélanges sur 65 données) dont la profondeur moyenne est supérieure à 10X (Mason et al, 2020). Leur étude présente aussi la mise au point d'un pipeline bioinformatique spécifique pour la détection des insertions rétrovirales, qu'il serait intéressant d'appliquer aux données du projet français.…”