Medicines and agricultural biocides are often discovered using large phenotypic screens across hundreds of compounds, where visible effects of whole organisms are compared to gauge efficacy and possible modes of action. However, such analysis is often limited to human-defined and static features. Here, we introduce a novel framework that can characterize shape changes (morphodynamics) for cell-drug interactions directly from images, and use it to interpret perturbed development of Phakopsora pachyrhizi, the Asian soybean rust crop pathogen. We describe population development over a 2D space of shapes (morphospace) using two models with condition-dependent parameters: a top-down Fokker-Planck model of diffusive development over Waddington-type landscapes, and a bottom-up model of tip growth. We discover a variety of landscapes, describing phenotype transitions during growth, and identify possible perturbations in the tip growth machinery that cause this variation. This demonstrates a widely-applicable integration of unsupervised learning and biophysical modeling.

No abstract

Computer assisted history matching has received much interest in the recent past. Several commercial programs are now available. We have tested some of these programs – most did not meet our expectations and we discontinued using them. The latest software tested helps in fine-tuning individual well behavior in a simulation model that has been globally matched. The program identifies model cells which have the highest influence on individual well behavior and modifies permeability in three directions (and porosity) such that the match is improved. Maximum allowed changes of permeability (and porosity) are controlled by the user. The procedure is repeated 10 to 20+ times until the match is considered satisfactory or the model no longer improves. A flow simulation model is globally matched when all wells can flow historical rates, the pressure history in different compartments or layers can be reproduced, external aquifer properties are adjusted, and fluid and rock properties are verified. In such a model computed well watercut can be grossly in error, some wells are too high others too low. Manual fine-tuning of individual well behavior is a lengthy and frustrating process. We have tested the new computer assisted history matching program on several models. Improvements to most well matches were spectacular. Model set up and history match procedures used are outlined on a moderately complex reservoir containing more than 1 billion STB oil initially in place. Production history spans 14 years, current recovery is ~25% of OOIP. Some 40 wells are active, four of them peripheral injectors. Injection voidage replacement is ~30%; there is a strong bottom and edge aquifer. The reservoir consists of 3 geological layers; the upper one has a permeability of several Darcy; the middle layer is highly heterogeneous and forms a partial seal to the lower layer with permeability of ~100 mD. We show examples of individual well matches before and after computer assisted fine tuning. We also discuss limitations of the program used and manual interferences required for selected wells.