Biology subjects

Kahraman, M.

Publications and source records attributed to Kahraman, M..

3 recordsLinked to original sources

bioRxiv2024

TransfoRNA: Navigating the Uncertainties of Small RNA Annotation with an Adaptive Machine Learning Strategy

Small RNAs hold crucial biological information and have immense diagnostic and therapeutic value. While many established annotation tools focus on microRNAs, there are myriads of other small RNAs that are currently underutilized. These small RNAs can be difficult to annotate, as ground truth is limited and well-established mapping and mismatch rules are lacking. TransfoRNA is a machine learning framework based on Transformers that explores an alternative strategy. It uses common annotation tools to generate a small seed of high-confidence training labels, while then expanding upon those labels iteratively. TransfoRNA learns sequence-specific representations of all RNAs to construct a similarity network which can be interrogated as new RNAs are annotated, allowing to rank RNAs based on their familiarity. While models can be flexibly trained on any RNA dataset, we here present a version trained on TCGA (The Cancer Genome Atlas) small RNA sequences and demonstrate its ability to add annotation confidence to an unrelated dataset, where 21% of previously unannotated RNAs could be annotated. Relative to its training data, TransfoRNA could boost high-confidence annotations in TCGA by [~]50% while providing transparent explanations even for low-confidence ones. It could learn to annotate 97% of isomiRs from just single examples and confidently identify new members of other familiar classes with high accuracy, while reliably rejecting false RNAs. All source code is available at https://github.com/gitHBDX/TransfoRNA and can be executed at Code Ocean (https://codeocean.com/capsule/5415298/). An interactive website is available at www.transforna.com. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=149 SRC="FIGDIR/small/599329v1_ufig1.gif" ALT="Figure 1"> View larger version (53K): org.highwire.dtl.DTLVardef@11bc3d9org.highwire.dtl.DTLVardef@1d6ff01org.highwire.dtl.DTLVardef@1ffac02org.highwire.dtl.DTLVardef@75ccd8_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2024

ctQC improves biological inferences from single cell and spatial transcriptomics data

Quality control (QC) is the first critical step in single cell and spatial data analysis pipelines. QC is particularly important when analysing data from primary human samples, since genuine biological signals can be obscured by debris, perforated cells, cell doublets and ambient RNA released into the "soup" by cell lysis. Consequently, several QC methods for single cell data, employ fixed or data-driven quality thresholds. While these approaches efficiently remove empty droplets, they often retain low-quality cells. Here, we propose cell type-specific QC (ctQC), a stringent, data-driven QC approach that adapts to cell type differences and discards soup and debris. Evaluating single cell RNA-seq data from colorectal tumors, human spleen, and peripheral blood mononuclear cells, we demonstrate that ctQC outperforms existing methods by improving cell type separation in downstream clustering, suppressing cell stress signatures, revealing patient-specific cell states, eliminating artefactual clusters and reducing ambient RNA artifacts. When applied to sequencing-based spatial RNA profiling data (Slide-seq), ctQC improved spatial coherence of cell clusters and consistency with anatomical structures. These results demonstrate that strict, data-driven, cell-type-specific QC is applicable to diverse sample types and substantially improves the quality and reliability of biological inferences from single cell and spatial RNA profiles.

bioinformatics↗

bioRxiv2020

Deep sncRNA-seq of the PPMI cohort to study Parkinson's disease progression

Coding and non-coding RNAs have diagnostic and prognostic importance in Parkinsons diseases (PD). We studied circulating small non-coding RNAs (sncRNAs) in 7, 003 samples from two longitudinal PD cohorts (Parkinsons Progression Marker Initiative (PPMI) and Luxembourg Parkinsons Study (NCER-PD)) and modelled their influence on the transcriptome. First, we sequenced sncRNAs in 5, 450 blood samples of 1, 614 individuals in PPMI. The majority of 323 billion reads (59 million reads per sample) mapped to miRNAs. Other covered RNA classes include piRNAs, rRNAs, snoRNAs, tRNAs, scaRNAs, and snRNAs. De-regulated miRNAs were associated with the disease and disease progression and occur in two distinct waves in the third and seventh decade of live. Originating mostly from a characteristic set of immune cells they resemble a systemic inflammation response and mitochondrial dysfunction, two hallmarks of PD. By profiling 1, 553 samples from 1, 024 individuals in the NCER-PD cohort using an independent technology, we validate relevant findings from the sequencing study. Finally, network analysis of sncRNAs and transcriptome sequencing of the original cohort identified regulatory modules emerging in progressing PD patients.

neuroscience↗