High throughput screening determines the effects of many conditions on a given biological target. Currently, to estimate the effects of those conditions on other targets requires either strong modeling assumptions (e.g. similarities among targets) or separate screens. Ideally, data-driven experimentation could be used to learn accurate models for many conditions and targets without doing all possible experiments. We have previously described an active machine learning algorithm that can iteratively choose small sets of experiments to learn models of multiple effects. We now show that, with no prior knowledge and with liquid handling robotics and automated microscopy under its control, this learner accurately learned the effects of 48 chemical compounds on the subcellular localization of 48 proteins while performing only 29% of all possible experiments. The results represent the first practical demonstration of the utility of active learning-driven biological experimentation in which the set of possible phenotypes is unknown in advance.
© 2016, Naik et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
DNA base lesions, such as incorporation of uracil into DNA or base mismatches, can be mutagenic and toxic to replicating cells. To discover factors in repair of genomic uracil, we performed a CRISPR knockout screen in the presence of floxuridine, a chemotherapeutic agent that incorporates uracil and fluorouracil into DNA. We identified known factors, such as uracil DNA N-glycosylase (UNG), and unknown factors, such as the N6-adenosine methyltransferase, METTL3, as required to overcome floxuridine-driven cytotoxicity. Visualized with immunofluorescence, the product of METTL3 activity, N6-methyladenosine, formed nuclear foci in cells treated with floxuridine. The observed N6-methyladenosine was embedded in DNA, called 6mA, and these results were confirmed using an orthogonal approach, liquid chromatography coupled to tandem mass spectrometry. METTL3 and 6mA were required for repair of lesions driven by additional base-damaging agents, including raltitrexed, gemcitabine, and hydroxyurea. Our results establish a role for METTL3 and 6mA in promoting genome stability in mammalian cells, especially in response to base damage.
Niemann–Pick disease type C (NPC) is a devastating lysosomal storage disease characterized by abnormal cholesterol accumulation in lysosomes. Currently, there is no treatment for NPC. Transcription factor EB (TFEB), a member of the microphthalmia transcription factors (MiTF), has emerged as a master regulator of lysosomal function and promoted the clearance of substrates stored in cells. However, it is not known whether TFEB plays a role in cholesterol clearance in NPC disease. Here, we show that transgenic overexpression of TFEB, but not TFE3 (another member of MiTF family) facilitates cholesterol clearance in various NPC1 cell models. Pharmacological activation of TFEB by sulforaphane (SFN), a previously identified natural small-molecule TFEB agonist by us, can dramatically ameliorate cholesterol accumulation in human and mouse NPC1 cell models. In NPC1 cells, SFN induces TFEB nuclear translocation via a ROS-Ca2+-calcineurin-dependent but MTOR-independent pathway and upregulates the expression of TFEB-downstream genes, promoting lysosomal exocytosis and biogenesis. While genetic inhibition of TFEB abolishes the cholesterol clearance and exocytosis effect by SFN. In the NPC1 mouse model, SFN dephosphorylates/activates TFEB in the brain and exhibits potent efficacy of rescuing the loss of Purkinje cells and body weight. Hence, pharmacological upregulating lysosome machinery via targeting TFEB represents a promising approach to treat NPC and related lysosomal storage diseases, and provides the possibility of TFEB agonists, that is, SFN as potential NPC therapeutic candidates.