Umeå University's logo

Saccharomyces cerevisiae is a widely used model organism for the molecular analysis of genes and proteins. Several methods have been developed to study protein function and activity through heterologous gene expression, including yeast two-hybrid and yeast complementation. Traditionally, these yeast-based assays were performed on solid agar plates. While this approach provides an easy visual readout, it is difficult to quantify the results accurately. To overcome this limitation, liquid-based methods were introduced. Most of these methods rely on the use of spectrophotometry to measure reduction in light transmission as a result of light scattering and monitor culture growth.

In this study, we propose nephelometry as an additional method for performing and analyzing liquid-culture yeast complementation assays. More specifically, we compare the suitability of using nephelometry for the functional analysis of two homologous proteins using yeast complementation: The amino acid transporter homologues Arabidopsis thaliana LYSINE HISTIDINE TRANSPORTER 1 (AtLHT1) and Populus tremula L. x tremuloides Michx LYSINE HISTIDINE TRANSPORTER 1.2 (PtrLHT1.2). In previous reports, no differences in microbial growth were detected, irrespective of which homolog was used to rescue an amino acid-deficient yeast mutant strain. By using nephelometry to record yeast growth, we demonstrated that it is a robust and reproducible method. When comparing to spectrophotometric measurements of yeast cultures, it proved to be a suitable alternative.

The novel approach even revealed previously undetected differences in culture growth of both homologues, highlighting nephelometry's potential to improve sensitivity in yeast-based functional assays.

We present the use of nephelometry as an equal method to yeast complementation traditionally executed on solid agar medium or in liquid culture with spectrophotometric analysis.

Background and Objective: The advancement of generative AI in medical imaging faces the trilemma of simultaneously achieving high fidelity and diversity in synthetic data generation. Although Generative Adversarial Networks (GANs) have demonstrated significant potential, they are often hindered by limitations such as mode collapse and poor coverage of real data distributions. This study investigates the use of GAN ensembles as a solution to these challenges, with the goal of enhancing the quality and utility of synthetic medical images.

Methods: We formulate a multi-objective optimisation problem to select an optimal ensemble of GANs that balances fidelity and diversity. The ensemble comprises models that contribute uniquely to the synthetic data space, ensuring minimal redundancy. A comprehensive evaluation was conducted using three distinct medical imaging datasets. We tested 22 GAN architectures, incorporating various loss functions and regularisation techniques. By sampling models at different training epochs, we crafted 110 unique configurations for ensemble selection.

Results: The selected GAN ensembles demonstrated improved performance in generating synthetic medical images that closely resemble real data distributions. These ensembles preserved image fidelity while increasing diversity. In some settings, downstream models trained on synthetic data achieved slightly higher accuracy than those trained on real data alone. This effect arises because the synthetic images act as a targeted data augmentation mechanism that enhances class balance and diversity rather than replacing real data.

Conclusions: GAN ensembles offer a robust solution to the fidelity–diversity–efficiency trade-off in medical image synthesis. By integrating multiple complementary models, the proposed approach improves the representativeness and utility of synthetic medical data, potentially advancing a wide range of clinical and research applications in diagnostic AI.

Background: The established method in primary care to distinguish skin cancer from benign lesions is clinical examination, with or without dermoscopy. The experience among primary care physicians in assessing skin tumours varies, as does the accessibility to teledermoscopy. To enhance diagnostic performance, improved methods for skin tumour assessment are warranted. The aim of this study was to investigate the diagnostic performance of a non-invasive method that combines near-infrared spectroscopy with skin impedance measurement (NIRIMP) to detect skin cancer in primary care.

Material and Methods: NIRIMP measurements were collected prospectively from patients seeking primary care for skin lesion examination. The measurements were compared to the true lesion diagnosis using several machine learning methods, to determine the best machine learning methods to use and to determine the diagnostic performance of NIRIMP in distinguishing skin cancer from benign lesions.

Results: Eighty participants with 109 lesions were included. Among these, 50 skin cancers or in situ cancers were detected: eight melanomas/in situ melanomas, four in situ squamous cell carcinomas, and 38 basal cell carcinomas. The ability of NIRIMP to distinguish any skin cancer/in situ cancer, as illustrated by the area under the receiver operating characteristics curve (AUC), was 0.776 and for melanomas/in situ melanomas alone the AUC was 0.911. When detecting any skin cancer, the AUC was slightly higher for NIR alone (0.826) compared to NIRIMP (0.776), whereas for IMP alone it was slightly lower (0.693).

Conclusions: Near infrared spectroscopy appears to be a promising bioengineering technique to detect skin cancer in primary care settings, of potential benefit for future skin lesion assessment. However, there was no compelling evidence supporting the benefit of adding skin impedance to improve diagnostic performance.

Multi-object grasping is a challenging task. It is important for energy and cost-efficient operation of industrial crane manipulators, such as those used to collect tree logs from the forest floor and on forest machines. In this work, we used synthetic data from physics simulations to explore how data-driven modeling can be used to infer multi-object grasp poses from images. We showed that convolutional neural networks can be trained specifically for synthesizing multi-object grasps. Using RGB-Depth images and instance segmentation masks as input, a U-Net model outputs grasp maps with the corresponding grapple orientation and opening width. Given an observation of a pile of logs, the model can be used to synthesize and rate the possible grasp poses and select the most suitable one, with the possibility to respect changing operational constraints such as lift capacity and reach. When tested in simulation on previously unseen data, the proposed model found successful grasp poses with an accuracy of up to 96%.

Machine learning models—deep neural networks in particular—have performed remarkably well on benchmark datasets across a wide variety of domains. However, the ease of finding adversarial counter-examples remains a persistent problem when training times are measured in hours or days and the time needed to find a successful adversarial counter-example is measured in seconds. Much work has gone into generating and defending against these adversarial counter-examples, however the relative costs of attacks and defences are rarely discussed. Additionally, machine learning research is almost entirely guided by test/train metrics, but these would require billions of samples to meet industry standards. The present work addresses the problem of understanding and predicting how particular model hyper-parameters influence the performance of a model in the presence of an adversary. The proposed approach uses survival models, worst-case examples, and a cost-aware analysis to precisely and accurately reject a particular model change during routine model training procedures rather than relying on real-world deployment, expensive formal verification methods, or accurate simulations of very complicated systems (e.g., digitally recreating every part of a car or a plane). Through an evaluation of many pre-processing techniques, adversarial counter-examples, and neural network configurations, the conclusion is that deeper models do offer marginal gains in survival times compared to more shallow counterparts. However, we show that those gains are driven more by the model inference time than inherent robustness properties. Using the proposed methodology, we show that ResNet is hopelessly insecure against even the simplest of white box attacks.

Predicting future interactions or novel links in networks is an indispensable tool across diverse domains, including genetic research, online social networks, and recommendation systems. Among the numerous techniques developed for link prediction, those leveraging the networks' community structure have proven highly effective. For example, the recently proposed MapSim predicts links based on a similarity measure derived from the code structure of the map equation, a community-detection objective function that operates on network flows. However, the standard map equation assumes complete observations and typically identifies many small modules in networks where the nodes connect through only a few links. This aspect can degrade MapSim's performance on sparse networks. To overcome this limitation, we propose to incorporate a global regularization method based on a Bayesian estimate of the transition rates along with three local regularization methods. The regularized versions of the map equation compensate for incomplete observations and mitigate spurious community fragmentation in sparse networks. The regularized methods outperform standard MapSim and several state-of-the-art embedding methods in highly sparse networks. This performance holds across multiple real-world networks with randomly removed links, simulating incomplete observations. Among the proposed regularization methods, the global approach provides the most reliable community detection and the highest link prediction performance across different network densities. The principled method requires no hyperparameter tuning and runs at least an order of magnitude faster than the embedding methods.

Purpose: To make Deep Image Prior (DIP)-based parameter mapping faster, more accurate, and suitable for clinical applications, with added support for multislice and 3D datasets.

Methods: DIP leverages the inherent structure of an untrained image generator to address various inverse imaging tasks, including denoising. In this study, we enhance DIP-based denoising for parameter mapping with warm-start across neighboring image slices and different patient subjects. This approach leverages spatial similarity to reduce computation time. Additionally, we introduce an early-stopping criterion that selects the denoising level based on MRI signal noise. We further investigate uncertainty calibration through dropout probability tuning to address issues with miscalibrated uncertainty estimates from Monte Carlo dropout. Furthermore, we explore reducing computation time by tuning learning rates and network complexity.

Results: We show that reusing image generator weights with warm-start significantly accelerates the denoising of large datasets, reducing computation time by 78% to 95% across various tasks. The early stopping approach proved effective, eliminating the need to manually select the number of optimization steps. Dropout probability tuning helps mitigate the issue of miscalibrated uncertainty, though further refinements are necessary, particularly to achieve better calibration on a per-pixel level. Additionally, tuning learning rates and network complexity provided valuable insights into optimizing the model for different tasks.

Conclusion: The proposed developments enable DIP-based parameter mapping to become faster, more accurate, and, consequently, more practical and scalable for clinical applications involving larger datasets.

The world of Machine Learning (ML) offers great promise, but often there is a noticeable gap between claims made in research papers and the model's practical performance in real-life applications. This gap can often be attributed to systematic errors and pitfalls that occur during the development phase of ML models. This study aims to systematically identify these errors. For this, we break down the ML process into four main stages: data handling, model design, model evaluation, and reporting. Across these stages, we have identified fourteen common pitfalls based on a comprehensive review of around 60 papers discussing either broad challenges or specific pitfalls within ML pipeline. Moreover, Using the Brain Tumor Segmentation (BraTS) dataset, we perform three experiments to illustrate the impacts of these pitfalls, providing examples of how they can skew results and affect outcomes. In addition, we also perform a review to study the frequency of unclear reporting regarding these pitfalls in ML research. The goal of this review was to assess whether authors have adequately addressed these pitfalls in their reports. For this, we review 126 randomly chosen papers on image segmentation from the ICCV (2013-2021) and MICCAI (2013-2022) conferences from the last ten years. The results from this review show a notable oversight of these issues, with many of the papers lacking clarity on how the pitfalls are handled. This highlights an important gap in current reporting practices within the ML community. The code for the experiments is available at https://github.com/SG-Azar/BraTS-ML-Pitfalls-Experiments.

Data Augmentation (DA) is a technique to increase the quantity and diversity of the training data, and by that alleviate overfitting and improve generalisation. However, standard DA produces synthetic data for augmentation with limited diversity. Generative Adversarial Networks (GANs) may unlock additional information in a dataset by generating synthetic samples having the appearance of real images. However, these models struggle to simultaneously address three key requirements: fidelity and high-quality samples; diversity and mode coverage; and fast sampling. Indeed, GANs generate high-quality samples rapidly, but have poor mode coverage, limiting their adoption in DA applications. We propose LatentAugment, a DA strategy that overcomes the low diversity of GANs, opening up for use in DA applications. Without external supervision, LatentAugment modifies latent vectors and moves them into latent space regions to maximise the synthetic images' diversity and fidelity. It is also agnostic to the dataset and the downstream task. A wide set of experiments shows that LatentAugment improves the generalisation of a deep model translating from MRI-to-CT beating both standard DA as well GAN-based sampling. We further demonstrate its effectiveness when translating from low-energy mammograms to dual-energy subtracted images in contrast-enhanced spectral mammography. Moreover, still in comparison with GAN-based sampling, LatentAugment synthetic samples show superior mode coverage and diversity. Code is available at: https://github.com/ltronchin/LatentAugment.