A DNA chip, a canopy model and an EEG pain study each map a pattern at one scale
A DNA-repair chip, a canopy model of soil fungi and an EEG pain study each report a biological pattern at the scale their design could hold.
Science··Morning
Three repair enzymes choose sequence context on a chip
A team at the Weizmann Institute of Science built a chip carrying thousands of short DNA molecules that hold the same lesion in different sequence surroundings, then measured how strongly fluorescently labelled repair enzymes bound to each one. The three glycosylases tested showed clear preferences: they chose targets by the letters around the damage rather than treating every damaged site as equal. The measured preferences line up with mutational signatures seen in human tumours, which the authors read as a sign that where repair works well or poorly helps shape where mutations accumulate in the genome. The correlation is between binding strength on a chip and patterns catalogued in tumour genomes, so the design stops short of showing that the enzymes' preferences cause those patterns in living tissue. The study appeared in Nature Communications. What the experiment measured is sequence-dependent binding under controlled chip conditions. The pattern holds at the scale of short synthetic molecules and mapped binding strength, as a laboratory binding map rather than a clinical map of repair inside patients.[1]
Canopy images predict fungal turnover under one plot
Wanwan Yu and colleagues at the University of Alberta paired 538 DNA-sequenced soil samples from a 26-hectare plot with drone imagery of the canopy above, then trained a random forest model to predict fungal diversity from what the camera could see. The plot lies in a 40-year-old planted forest inside a nature reserve in China. The drone data covered tree height and leaf reflectance, which stands in for chlorophyll and water content. The model accounted for about 53 per cent of beta diversity, the turnover in fungal community composition from place to place, and between 28 per cent and 45 per cent of alpha diversity, the richness within a single patch, depending on the measure used. The gap between the two figures matters for anyone hoping to replace soil sampling with aerial survey: the composition signal travels upward more reliably than the count of species does. The paper appeared in Forest Ecology and Management. The pattern therefore holds for one subtropical planted plot and for diversity statistics the model could learn from canopy structure. It is a prediction skill under those sampling and imaging conditions, reported for that plot rather than as a reading of every forest from the air.[2]
Pain markers hold within a person, across cohorts only for within-person swings
Laura Tiemann and colleagues recorded electroencephalography from 161 healthy participants and repeated the analysis in a separate cohort of 111. Neural patterns tied to a single person's fluctuating pain held up both over time and across the two cohorts. The patterns linked to differences between people behaved otherwise: they were repeatable over time within the first cohort but did not replicate in the second. The authors read the split as evidence that within-person and between-person variability in pain perception are encoded by distinct neural patterns that differ in robustness, which points brain-based markers towards tracking how one patient's pain moves rather than towards ranking one patient against another. The study is an observational analysis of healthy volunteers under experimental stimulation, so it leaves clinical pain conditions untested. It was published in PLOS Biology. Read beside the chip and canopy studies, the common thread is scale. Sequence preference appears on short DNA molecules, fungal turnover under one imaged canopy, and pain encoding inside repeated measures of the same people. Each pattern is reported where the measurement design could hold it.[3]
