The origins of unpredictability in life outcome prediction tasks.

Why are some life outcomes difficult to predict? We investigated this question through in-depth qualitative interviews with 40 families sampled from a multidecade longitudinal study. Our sampling and interviewing process was informed by the earlier efforts of hundreds of researchers to predict life outcomes for participants in this study. The qualitative evidence we uncovered in these interviews combined with a mathematical decomposition of prediction error led us to create a conceptual framework. Our specific evidence and our more general framework suggest that unpredictability should be expected in many life outcome prediction tasks, even in the presence of complex algorithms and large datasets. Our work provides a foundation for future empirical and theoretical work on unpredictability in human lives.

Srs2 binding to PCNA and its sumoylation contribute to RPA antagonism during the DNA damage response.

Activation of the DNA damage checkpoint upon genotoxin treatment induces a multitude of cellular changes, such as cell cycle arrest, to cope with genome stress. After prolonged genotoxin treatment, the checkpoint can be downregulated to allow cell cycle and growth resumption. In yeast, downregulation of the DNA damage checkpoint requires the Srs2 DNA helicase, which removes the ssDNA binding complex RPA and the associated Mec1 checkpoint kinase from DNA, thus dampening Mec1 activation. However, it is unclear whether the 'anti-checkpoint' role of Srs2 is temporally and spatially regulated to both allow timely checkpoint termination and to prevent superfluous RPA removal. Here we address this question by examining regulatory elements of Srs2, including its phosphorylation, sumoylation, and protein-interaction sites. Our genetic analyses and checkpoint level assessment suggest that the RPA countering role of Srs2 is promoted by Srs2 binding to PCNA, which is known to recruit Srs2 to subsets of ssDNA regions. RPA antagonism is further fostered by Srs2 sumoylation, which we found depends on the Srs2-PCNA interaction. Srs2 sumoylation is additionally reliant on Mec1 and peaks after Mec1 activity reaches maximal levels. Collectively, our data provide evidence for a two-step model wherein checkpoint downregulation is facilitated by PCNA-mediated Srs2 recruitment to ssDNA-RPA filaments and the subsequent Srs2 sumoylation stimulated upon Mec1 hyperactivation. We propose that this mechanism allows Mec1 hyperactivation to trigger checkpoint recovery.