Back

LMM-MQM time series mapping - An application in a murine advanced intercross line identifies novel growth QTLs

Arends, D.; Hesse, D.; Kärst, S.; Heise, S.; Lyu, S.; Korkuc, P.; Delpero, M.; Mulligan, M. K.; Prins, P.; Brockmann, G. A.

2022-01-25 genetics
10.1101/2022.01.23.477441 bioRxiv
Show abstract

The Berlin Fat Mouse Inbred line 860 (BFMI860) is a mouse model for juvenile obesity. Previously, a recessive major effect locus (jObes1) was identified on chromosome 3 explaining around 26% of the body weight variance in an BFMI860xC57BL/6NCrl advanced intercross line. The aim of this study was to discover additional QTL. Time series body weight data were modeled using linear mixed models (LMM), while a multiple QTL mapping (MQM) approach compensated for the jObes1 locus effect. LMM-MQM identified five additional loci significantly associated with body weight. Variance explained by the jObes1 locus increased to 38.1% when using LMM-MQM mapping, while the additional loci explained between 2.0% and 3.9% of the body weight variance. Several positional candidate genes within the novel QTL regions were found in KEGG pathways for insulin signaling and insulin resistance. Strong distortion with preference for the BFMI allele was observed within a newly identified QTL containing the well-known Foxo1 regulator of adipocyte differentiation. Here, we present a novel method for QTL detection in time series data: LMM-MQM time series mapping. We show that our method is more powerful in detecting QTLs compared to single timepoint mapping approaches. Thus, the time series structure should be considered for optimal detection of small effect QTLs. LMM-MQM time series mapping can be used to find genetic determinants of all kind of "phenotypes over time" be it lactation curves in cattle, plant biomass, drug clearance in human clinical trials, or cognitive decline during disease.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.