Sliding Window INteraction Grammar (SWING): a generalized interaction language model for peptide and protein interactions.

The explosion of sequence data has allowed the rapid growth of protein language models (pLMs). pLMs have now been employed in many frameworks including variant-effect and peptide-specificity prediction. Traditionally, for protein-protein or peptide-protein interactions (PPIs), corresponding sequences are either co-embedded followed by post-hoc integration or the sequences are concatenated prior to embedding. Interestingly, no method utilizes a language representation of the interaction itself. We developed an interaction LM (iLM), which uses a novel language to represent interactions between protein/peptide sequences. Sliding Window Interaction Grammar (SWING) leverages differences in amino acid properties to generate an interaction vocabulary. This vocabulary is the input into a LM followed by a supervised prediction step where the LM's representations are used as features. SWING was first applied to predicting peptide:MHC (pMHC) interactions. SWING was not only successful at generating Class I and Class II models that have comparable prediction to state-of-the-art approaches, but the unique Mixed Class model was also successful at jointly predicting both classes. Further, the SWING model trained only on Class I alleles was predictive for Class II, a complex prediction task not attempted by any existing approach. For de novo data, using only Class I or Class II data, SWING also accurately predicted Class II pMHC interactions in murine models of SLE (MRL/lpr model) and T1D (NOD model), that were validated experimentally. To further evaluate SWING's generalizability, we tested its ability to predict the disruption of specific protein-protein interactions by missense mutations. Although modern methods like AlphaMissense and ESM1b can predict interfaces and variant effects/pathogenicity per mutation, they are unable to predict interaction-specific disruptions. SWING was successful at accurately predicting the impact of both Mendelian mutations and population variants on PPIs. This is the first generalizable approach that can accurately predict interaction-specific disruptions by missense mutations with only sequence information. Overall, SWING is a first-in-class generalizable zero-shot iLM that learns the language of PPIs.

SLIDE: Significant Latent Factor Interaction Discovery and Exploration across biological domains.

Modern multiomic technologies can generate deep multiscale profiles. However, differences in data modalities, multicollinearity of the data, and large numbers of irrelevant features make analyses and integration of high-dimensional omic datasets challenging. Here we present Significant Latent Factor Interaction Discovery and Exploration (SLIDE), a first-in-class interpretable machine learning technique for identifying significant interacting latent factors underlying outcomes of interest from high-dimensional omic datasets. SLIDE makes no assumptions regarding data-generating mechanisms, comes with theoretical guarantees regarding identifiability of the latent factors/corresponding inference, and has rigorous false discovery rate control. Using SLIDE on single-cell and spatial omic datasets, we uncovered significant interacting latent factors underlying a range of molecular, cellular and organismal phenotypes. SLIDE outperforms/performs at least as well as a wide range of state-of-the-art approaches, including other latent factor approaches. More importantly, it provides biological inference beyond prediction that other methods do not afford. Thus, SLIDE is a versatile engine for biological discovery from modern multiomic datasets.

A supervised take on dimensionality reduction via hybrid subset selection.

Amouzgar et al. present HSS-LDA, a supervised dimensionality reduction approach for single-cell data that outperforms existing unsupervised techniques. They couple hybrid subset selection to linear discriminant analysis and identify interpretable linear combinations of predictors that best separate predefined biological groups.

DataRemix: a universal data transformation for optimal inference from gene expression datasets.

MOTIVATION: RNA-seq technology provides unprecedented power in the assessment of the transcription abundance and can be used to perform a variety of downstream tasks such as inference of gene-correlation network and eQTL discovery. However, raw gene expression values have to be normalized for nuisance biological variation and technical covariates, and different normalization strategies can lead to dramatically different results in the downstream study. RESULTS: We describe a generalization of singular value decomposition-based reconstruction for which the common techniques of whitening, rank-k approximation and removing the top k principal components are special cases. Our simple three-parameter transformation, DataRemix, can be tuned to reweigh the contribution of hidden factors and reveal otherwise hidden biological signals. In particular, we demonstrate that the method can effectively prioritize biological signals over noise without leveraging external dataset-specific knowledge, and can outperform normalization methods that make explicit use of known technical factors. We also show that DataRemix can be efficiently optimized via Thompson sampling approach, which makes it feasible for computationally expensive objectives such as eQTL analysis. Finally, we apply our method to the Religious Orders Study and Memory and Aging Project dataset, and we report what to our knowledge is the first replicable trans-eQTL effect in human brain. AVAILABILITYAND IMPLEMENTATION: DataRemix is an R package which is freely available at GitHub (https://github.com/wgmao/DataRemix). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Predictors of time to first cannulation for arteriovenous fistula in pediatric hemodialysis patients: Midwest Pediatric Nephrology Consortium study.

BACKGROUND: Permanent vascular access (PVA) is preferred for long-term hemodialysis. Arteriovenous fistulae (AVF) have the best patency and the lowest complication rates compared to arteriovenous grafts (AVG) and tunneled cuffed catheters (TCC). However, AVF need time to mature. This study aimed to investigate predictors of time to first cannulation for AVF in pediatric hemodialysis patients. METHODS: Data on first AVF and AVG of patients at 20 pediatric dialysis centers were collected retrospectively, including demographics, clinical information, dialysis markers, and surgical data. Statistical modeling was used to investigate predictors of outcome. RESULTS: First PVA was created in 117 children: 103 (88%) AVF and 14 (12%) AVG. Mean age at AVF creation was 15.0 ± 3.3 years. AVF successfully matured in 89 children (86.4%), and mean time to first cannulation was 3.6 ± 2.5 months. In a multivariable regression model, study center, age, duration of non-permanent vascular access (NPVA), and Kt/V at AVF creation predicted time to first cannulation, with study center as the strongest predictor (p < 0.01). Time to first cannulation decreased with increasing age (p = 0.03) and with increasing Kt/V (p = 0.01), and increased with duration of NPVA (p = 0.03). Secondary failure occurred in 10 AVF (11.8%). Time to first cannulation did not predict secondary failure (p = 0.29), but longer time to first cannulation tended towards longer secondary patency (p = 0.06). CONCLUSIONS: Study center is the strongest predictor of time to first cannulation for AVF and deserves further investigation. Time to first cannulation is significantly shorter in older children, with more efficient dialysis treatments, and increases with longer NPVA duration.

Correction to: Predictors of patency for arteriovenous fistulae and grafts in pediatric hemodialysis patients.

The original version of this article unfortunately contained a mistake. The name of Vimal Chadha was presented incorrectly. The corrected author list is given above.

Predictors of patency for arteriovenous fistulae and grafts in pediatric hemodialysis patients.

BACKGROUND: Hemodialysis (HD) guidelines recommend permanent vascular access (PVA) in children unlikely to receive kidney transplant within 1 year of starting HD. We aimed to determine predictors of primary and secondary patency of PVA in pediatric HD patients. METHODS: Retrospective chart reviews were performed for first PVAs in 20 participating centers. Variables collected included patient demographics, complications, interventions, and final outcome. RESULTS: There were 103 arterio-venous fistulae (AVF) and 14 AV grafts (AVG). AVF demonstrated superior primary (p = 0.0391) and secondary patency (p = 0.0227) compared to AVG. Primary failure occurred in 16 PVA (13.6%) and secondary failure in 14 PVA (12.2%). AVF were more likely to have primary failure (odds ratio (OR) = 2.10) and AVG had more secondary failure (OR = 3.33). No demographic, clinical, or laboratory variable predicted primary failure of PVA. Anatomical location of PVA was predictive of secondary failure, with radial having the lowest risk compared to brachial (OR = 12.425) or femoral PVA (OR = 118.618). Intervention-free survival was predictive of secondary patency for all PVA (p = 0.0252) and directly correlated with overall survival of AVF (p = 0.0197) but not AVG. Study center demonstrated statistically significant effect only on intervention-free AVF survival (p = 0.0082), but not number of complications or interventions, or outcomes. CONCLUSIONS: In this multi-center pediatric HD cohort, AVF demonstrated primary and secondary patency advantages over AVG. Radial PVA was least likely to develop secondary failure. Intervention-free survival was the only predictor of secondary patency for AVF and directly correlated with overall access survival. The study center effect on intervention-free survival of AVF deserves further investigation.