Final Paper of Bioinformatics Student http://http://repository.i3l.ac.id:80/jspui/handle/123456789/88 2026-04-21T21:59:36Z 2026-04-21T21:59:36Z Austin, Felicia http://http://repository.i3l.ac.id:80/jspui/handle/123456789/1293 2026-01-12T04:56:46Z 2025-08-31T00:00:00Z

Title: Developing Polygenic Risk Score Pipeline for Obesity in Indonesia Using Nextflow Authors: Austin, Felicia Abstract: Obesity has become one of the biggest public health problems in the world, including in Indonesia (Haththotuwa et al., 2020). One such way to potentially prevent obesity is through a gene prediction method called polygenic risk score (PRS). PRS uses genomic and phenotypic data to determine which gene variants have a higher or lower potential to develop a certain disease (Lewis & Vassos, 2020). Pipeline (also known as a workflow) in bioinformatics refers to the steps of an analysis where the output of a step becomes the input for the next, until it achieves the end result (Leipzig, 2017). As usual with technologies, the tools and methods of a pipeline are ever evolving with the discovery of better algorithms and computational capabilities (Wratten et al., 2019). Combined with the fact that even a simple PRS analysis can contain multiple different softwares, the importance of a standardized and reproducible pipeline becomes an urgent concern. Hence, the aim of this paper is to create a pipeline in Nextflow to calculate obesity PRS for Indonesian patients using the clumping and thresholding method. The pipeline included initial data cleaning, merging, and analysis, PRS calculation using clumping and thresholding method, and result visualizations (which include ROC curve and bell curve for distribution analysis).

2025-08-31T00:00:00Z Lunoto, Dennis http://http://repository.i3l.ac.id:80/jspui/handle/123456789/1292 2026-01-12T04:47:58Z 2025-08-31T00:00:00Z

Title: Analysis of Blood Cell of Acute Lymphocytic Leukemia Using Different Image Classification Models Authors: Lunoto, Dennis Abstract: Acute Lymphocytic Leukemia (ALL) is one of the most common and malignant tumors that is prevalent in both the young and old. However, it is most commonly found in very young children. Hence, early diagnosis and prognosis are important to improve the patient’s survival rate. With AI, improving screening, which is one of the methods of detection that is more sensitive and accurate will allow better analysis and will in turn assist medical professionals in their line of work. This research based on image classification will explore a deep learning-based model on convolutional neural networks (CNN), using four EfficientNet models including EfficientNetB0, EfficientNetB1, EfficientNetB2, and EfficietNetB3, to compare which of these models will produce the best results.

2025-08-31T00:00:00Z Heerlie, Devita Mayanda http://http://repository.i3l.ac.id:80/jspui/handle/123456789/1291 2026-01-12T04:43:15Z 2025-08-31T00:00:00Z

Title: A Pangenome Perspective on Genetic Diversity and Functional Potential of Giant Virus (Nucleocytoviricota) Authors: Heerlie, Devita Mayanda Abstract: Giant viruses belong to the phylum Nucleocytoviricota, a diverse group that comprises both isolated genomes (IGs) and giant virus metagenome-assembled genomes (GVMAGs). The diversity and gene content of giant viruses were analyzed using pangenome and gene-sharing network analyses. A large number of GVMAGs were recovered from marine environments through binning-based assembly methods, with IM_01 (Mesomimiviridae) representing the most abundant group. In contrast, recovery from soil environments was limited due to fragmented data. No environment-based patterns or clusters were observed in the phylogenetic tree. Functional annotation of the predicted genes revealed that most are involved in host metabolism and viral replication, with some unique functions, such as photosynthesis and multidrug resistance, also being identified. Pangenome analysis revealed varying levels of diversity across families, with some forming monophyletic groups and others exhibiting broader variation. Many GVMAGs lacking isolate genome representatives did not contain core genes, highlighting the importance of IGs in genome completeness and the need for taxonomic refinement. These patterns were often linked to host specificity. Gene interaction networks revealed distinct clustering across orders, further supporting the lineage-specific evolution of giant viruses.

2025-08-31T00:00:00Z Putta, Dhannyo http://http://repository.i3l.ac.id:80/jspui/handle/123456789/1290 2026-01-12T04:36:15Z 2025-08-31T00:00:00Z

Title: CNN-RNN Deep Learning Approach for ECG-Based Left Ventricular Ejection Fraction Prediction Authors: Putta, Dhannyo Abstract: Left ventricular ejection fraction (LVEF) has been used as prognostic indicator for heart failure and several other cardiovascular diseases. Current LVEF assessment relies on relatively inaccessible modalities such as transthoracic echocardiography (TTE), cardiac magnetic resonance (CMR), and computed tomography (CT). This study developed a CNN-RNN machine learning framework to predict LVEF from 5-seconds Holter and high-resolution ECG signals. Data used for training was obtained from the MUSIC dataset hosted on PhysioNet. Two types of deep learning architectures, such as CNN-LSTM, and CNN-GRU were trained and evaluated using mean average error (MAE), mean squared error (MSE), and Pearson correlation coefficient (PCC). The CNN-GRU model outperformed CNN-LSTM model. CNN-GRU model trained with high-resolution ECG reached PCC of 0.68 with MAE of 7.82 and MSE of 99.81. These values indicated a moderately strong positive correlation between predicted and actual LVEF values.

2025-08-31T00:00:00Z