gersteinlab.bsky.social
@gersteinlab.bsky.social
New @natcomms.nature.com ‬ paper led by @beaborsari.bsky.social ‬ & Mor Frank. Also thanks to co-authors Eve Wattenberg, @xuke0828.bsky.social , Susanna Liu, Xuezhu Yu & @markgerstein.bsky.social
August 25, 2025 at 3:38 PM
3/3 The test yielded a p-value of 4.91 × 10⁻⁸, which is far below the conventional significance threshold of 0.05. This indicates a statistically significant deviation in personality type distribution within the lab.
July 14, 2025 at 11:44 PM
2/3 In contrast, within the Gerstein Lab, there are 26 Analysts, 20 Diplomats, 6 Sentinels, and 4 Explorers. A chi-square goodness-of-fit test was conducted to evaluate whether the MBTI distribution in the lab significantly differs from that of the general population.
July 14, 2025 at 11:44 PM
1/3 Based on a survey of 22,678,145 individuals in the United States, the distribution of MBTI personality types in the general population is as follows: Analysts account for 16.72%, Diplomats for 44.43%, Sentinels for 23.91%, and Explorers for 14.93%
July 14, 2025 at 11:43 PM
4/4 ⚡ WANet + WALoss ⇒ 18 % faster SCF convergence & 1 000 × energy-error reduction vs. SOTA. One model, many properties—HOMO/LUMO, dipoles, electron densities—all from a single predicted Hamiltonian.
May 5, 2025 at 5:18 PM
3/4 🗂️ First release of PubChemQH—50 k large-molecule Hamiltonians (40–100 atoms) for robust benchmarking, generated by 128 GPUs for one month of processing, which motivates a scaling challenge which we refer to as SAD.
May 5, 2025 at 5:17 PM
2/4 🧩 We introduce Wavefunction-Alignment Loss + WANet, slashing SCF iterations while keeping ab-initio precision for molecules 3× larger than training data.
May 5, 2025 at 5:17 PM
Also thanks Y Li, S Liu, Y Gao, X Xin, S Lou, M Jensen, D Garrido, T Verplaetse, G Ash, J Zhang, M Girgenti, W Roberts, YaleCBB, YaleMBB, YaleBIDS, YalePsych, YaleCSDept, YaleData, UCIrvine, UniBarcelona, NIMHgov
December 19, 2024 at 6:11 PM
We show wearable-derived digital phenotypes improve accuracy of predicting adolescents affected by psychiatric disorders using AI models for time series. This enables continuous GWAS to identify genetic variants missed by traditional case-control GWAS.
www.youtube.com/watch?v=3Gv-...
Digital phenotyping using AI for psychiatric disorders and genetics
YouTube video by Jason Liu
www.youtube.com
December 19, 2024 at 6:10 PM