Divergence Time Estimation, a methodology that combines fossil record and molecular data, is how biologists time travel. But like the fictional time travel the biological time travel is extremely sensitive to parameter choice (including fossils used).I have been studying the local and global effects of fossils using an empirical data set, insect transcriptome phylogeny generated by the 1KITE (One Thousand Insect Transcriptome Evolution) consortium. My work shows that fossil calibrations can have a very strong effect on the estimated age of the node that they calibrate, while the presence of other calibrations throughout the tree might have minimal effect on neighboring nodes in the tree. It seems that the effect of the fossil calibrations is not additive as I had hypothesized but is rather localized. This is particularly relevant for older or deeper nodes, which were highly impacted by both local inclusion and exclusion of a calibration.
Since fossils heavily influences the result of divergence time estimation, we have to be very thorough in our fossil choice. To address this NESCent (National Evolutionary Synthesis Center) Fossil Calibration Working Group put together best-practice guidelines to choose fossils and established the Fossil Calibration Database. As part of this effort, I worked on a list of carefully chosen dragonfly fossils to be used for divergence time estimation. I have also worked on a similar list of vetted fossils for cockroaches as well. I have since used these fossils in understanding the evolutionary history of both dragonflies (in prep) and cockroaches (published).