| 1. |
- Kensert, Alexander, et al.
(författare)
-
Transfer Learning with Deep Convolutional Neural Networks for Classifying Cellular Morphological Changes
- 2019
-
Ingår i: SLAS Discovery. - : Elsevier BV. - 2472-5560 .- 2472-5552. ; 24:4, s. 466-475
-
Tidskriftsartikel (refereegranskat)abstract
- The quantification and identification of cellular phenotypes from high-content microscopy images has proven to be very useful for understanding biological activity in response to different drug treatments. The traditional approach has been to use classical image analysis to quantify changes in cell morphology, which requires several nontrivial and independent analysis steps. Recently, convolutional neural networks have emerged as a compelling alternative, offering good predictive performance and the possibility to replace traditional workflows with a single network architecture. In this study, we applied the pretrained deep convolutional neural networks ResNet50, InceptionV3, and InceptionResnetV2 to predict cell mechanisms of action in response to chemical perturbations for two cell profiling datasets from the Broad Bioimage Benchmark Collection. These networks were pretrained on ImageNet, enabling much quicker model training. We obtain higher predictive accuracy than previously reported, between 95% and 97%. The ability to quickly and accurately distinguish between different cell morphologies from a scarce amount of labeled data illustrates the combined benefit of transfer learning and deep convolutional neural networks for interrogating cell-based images.
|
|
| 2. |
- Siva, Kavitha, et al.
(författare)
-
A Phenotypic Screening Assay Identifies Modulators of Diamond Blackfan Anemia
- 2019
-
Ingår i: SLAS Discovery. - : Elsevier BV. - 2472-5552. ; 24:3, s. 304-313
-
Tidskriftsartikel (refereegranskat)abstract
- Diamond-Blackfan anemia (DBA) is a bone marrow failure syndrome caused by mutations in ribosomal protein genes. Pathogenic mechanisms are poorly understood but involve severely reduced proliferation of erythroid precursors. Because current DBA therapies are ineffective and associated with severe side effects, disease-specific therapies are urgently needed. We hypothesized that druggable molecular pathways underlying the defect can be revealed through phenotypic small-molecule screens. Accordingly, a screening assay was developed using c-kit+ fetal liver erythroid progenitors from a doxycycline-inducible DBA mouse model. The addition of doxycycline to the culture medium induces the phenotype and reduces proliferation to <10% of normal, such that rescue of proliferation can be used as a simple readout for screening. Here, we describe the assay rationale and efforts toward validation of a microtiter plate-compatible assay and its application in a pilot screen of 3871 annotated compounds. Ten hits demonstrated concentration-dependent activity, and we report a brief follow-up of one of these compounds. In conclusion, we established a robust scalable assay for screening molecules that rescue erythropoiesis in DBA.
|
|
