Topic
Sequencing and computational methods that measure gene expression in individual cells, used to map cell types, states, and marker programmes across tissues.
University of Sydney researchers used more than 14,000 human embryo cells to benchmark four leading blastoid methods and found large differences in fidelity. None fully captures early human development, but labs now have a shared molecular reference to test protocols against.
Reality
University of Sydney researchers scored four human blastoid methods against RNA from 14,000-plus embryo cells, and none fully matched a real blastocyst. Because the gaps differ by protocol, labs need to know which model produced a result before reading it as human biology.
Reality
Researchers testing 18 metrics on 14 datasets found MSE and Pearson delta often miscalibrated, too insensitive to register real skill in perturbation models. That weakens recent verdicts that the models do no better than an average guess.
Reality
A UC Berkeley team kept stem-cell-derived organoids in culture for up to a year and found TSC2-mutant astrocytes reactive from the moment they appeared. Helen Bateup says neurons may still contribute to the seizures.
Reality
Ninety-four models trained on more than 6 million nuclei from 1,494 prefrontal cortex donors assign risk genes for 12 brain disorders to neuronal, glial and immune populations. The signals hold in a veteran cohort of about 600,000.
Reality
TranscriptFormer learned gene expression the way a language model learns text, across 12 species from single-celled yeast to humans. Its cross-species matches arrive as hypotheses a bench experiment still has to test.
Reality
The method swaps curve-fitting for a Lomb-Scargle periodogram, so irregular pseudotime and branching trajectories need no branch assignments, at the cost of a ranked gene list that cannot say which lineage diverged.
Reality
Queries that once meant downloading petabytes of raw files now run in seconds against a reference-free index, which also makes isoform-level and off-reference variant questions askable rather than merely faster.
Reality
The first human evidence for the neurogenesis hypothesis arrives with a hedge from the lead author and no described way to measure the marker in a living patient.
Reality
A PNAS paper claims the first identification of the circuits that make cell-state changes stick. The validation so far is recovering fate determinants that biology already knew.
Reality
A University of Basel group has published Bonsai in Nature Biotechnology, replacing the two-dimensional map with a branching tree. The fidelity gains are asserted qualitatively, not quantified.
Reality