As proteomics experiments grow in scale, poor-quality measurements are increasingly hard to detect yet undermine biological conclusions. Here, the authors introduce MSstats + , a method that uses peak quality metrics and their temporal profiles to downweight poor measurements in differential analysis.