Benchmark of Data Processing Methods and Machine Learning Models for Gut Microbiome-Based Diagnosis of Inflammatory Bowel Disease

Patients with inflammatory bowel disease (IBD) wait months and undergo numerous invasive procedures between the initial appearance of symptoms and receiving a diagnosis. In order to reduce time until diagnosis and improve patient wellbeing, machine learning algorithms capable of diagnosing IBD from the gut microbiome's composition are currently being explored. To date, these models have had limited clinical application due to decreased performance when applied to a new cohort of patient samples. Various methods have been developed to analyze microbiome data which may improve the generalizability of machine learning IBD diagnostic tests. With an abundance of methods, there is a need to benchmark the performance and generalizability of various machine learning pipelines (from data processing to training a machine learning model) for microbiome-based IBD diagnostic tools. We collected fifteen 16S rRNA microbiome datasets (7,707 samples) from North America to benchmark combinations of gut microbiome features, data normalization and transformation methods, batch effect correction methods, and machine learning models. Pipeline generalizability to new cohorts of patients was evaluated with two binary classification metrics following leave-one-dataset-out cross (LODO) validation, where all samples from one study were left out of the training set and tested upon. We demonstrate that taxonomic features processed with a compositional transformation method and batch effect correction with the naive zero-centering method attain the best classification performance. In addition, machine learning models that identify non-linear decision boundaries between labels are more generalizable than those that are linearly constrained. Lastly, we illustrate the importance of generating a curated training dataset to ensure similar performance across patient demographics. These findings will help improve the generalizability of machine learning models as we move towards non-invasive diagnostic and disease management tools for patients with IBD.

Templated co-assembly into nanorods of polyanions and artificial virus capsid proteins

Recombinant triblock polypeptides C-S_n-B, where C is a 400 amino acid long hydrophilic random coil block, S_n is a multimer of the silk-like octapeptide S = (GAGAGAGQ), and B = K₁₂ is an oligolysine, have previously been shown to encapsulate double stranded DNA into rod-shaped, virus-like particles. In order to gain insight of the co-assembly process, and in order to be able to use these proteins for templating other types of nanorods, we here explore their co-assembly with a range of polyanionic templates: poly(acrylic acids) (PAA) of a wide range of lengths, poly(styrene sulphonate) (PSS) and the stiff anionic polysaccharide xanthan. The formation of the complexes was characterized using Dynamic Light Scattering (DLS), cryogenic Transmission Electronic Microscopy (Cryo-TEM) and Atomic Force Microscopy (AFM). Except at very high molar masses, we find that flexible anionic PAA and PSS lead to co-assembly of proteins with single polyanion chains into nanorods, with a packing factor as expected on the basis of charge stochiometry. Only for very long PAA templates (8 × 10⁵ Da) we find evidence for heterogeneous complexes with thin and thick sections. For the very stiff xanthan chains, we find that its stiffness precludes co-assembly with the artificial viral capsid proteins into condensed and regular nanorods. Given the simple and robust formation of rod-like structures with a range of polyanionic templates, we anticipate that the artificial virus proteins will be useful for preparing high-aspect ratio nanoparticles and scaffolds of precise size and find applications in nanotechnology and materials science for which currently natural rod-like viruses are being explored.

Gender differences between the effect of monosodium glutamate on food intake in rats

Monosodium glutamate (MSG) was recently reported to increase feeding in ad lib. fed rats [Reddy et al.(4)]. In the present study the responses of male and female adult rats to subcutaneous (SC) administration of MSG were compared. In female rats MSG 3 g/kg did not significantly affect food or water intake as compared with SC NaCl 0.9%; MSG 6 g/kg significantly reduced food intake but increased water intake. On the contrary, in male rats MSG 6 g/kg significantly increased food and water intake. It is concluded that there are differences between genders of rats in what concerns the effect of glutamate on food intake.

Diversifying Design of Nucleic Acid Aptamers Using Unsupervised Machine Learning

Inverse design of short single-stranded RNA and DNA sequences (aptamers) is the task of finding sequences that satisfy a set of desired criteria. Relevant criteria may be, for example, the presence of specific folding motifs, binding to molecular ligands, sensing properties, and so on. Most practical approaches to aptamer design identify a small set of promising candidate sequences using high-throughput experiments (e.g., SELEX) and then optimize performance by introducing only minor modifications to the empirically found candidates. Sequences that possess the desired properties but differ drastically in chemical composition will add diversity to the search space and facilitate the discovery of useful nucleic acid aptamers. Systematic diversification protocols are needed. Here we propose to use an unsupervised machine learning model known as the Potts model to discover new, useful sequences with controllable sequence diversity. We start by training a Potts model using the maximum entropy principle on a small set of empirically identified sequences unified by a common feature. To generate new candidate sequences with a controllable degree of diversity, we take advantage of the model's spectral feature: an "energy" bandgap separating sequences that are similar to the training set from those that are distinct. By controlling the Potts energy range that is sampled, we generate sequences that are distinct from the training set yet still likely to have the encoded features. To demonstrate performance, we apply our approach to design diverse pools of sequences with specified secondary structure motifs in 30-mer RNA and DNA aptamers.