Rucha Narkhede

A highly skilled and curious bioinformatics scientist with a strong background in research and thorough attention to detail, always eager to examine and work with complicated material. I'm actively looking for a work position in an organisation where I can use my skills of computer science and molecular biology.

• Benchmarking of Viral Quasispecies Assembly Algorithms (2022) Using realistic evolutionary viral populations, we assessed six viral quasispecies assembly tools on various performance metrics. Introduced a new high quality benchmarking dataset for viral quasispecies assembly from real samples with different complexities (eg. different virus mixes with varying coverages). The sample was obtained from Harvard Lab. A de novo method(VG-Flow) is used with algorithmic tweaks for reconstructing full-length haplotypes from pre-assembled contigs of challenging mixed samples where an accuracy of 96% was achieved. Classify ‘Kostic’ Microbiome data into healthy & tumor using Machine Learning (2021) Stacked Emsemble Model of Logistic Regression, KNN and Random Forest. Obtained an accuracy of ~74% and also found comparable results of phyla contributing to PC1 variation. Linear and Non-Linear Models to Classify Single-Cell Transcriptomics of Chronic Myeloid Leukemia Patients During Diagnosis and Remission Therapy Development (2021) The purpose of this project was to classify HSC, BCR-ABL+ Diagnosis and BCR-ABL+ Remission using various machine learning techniques. Gender Detection by Image and Voice Recognition using Deep Learning (2019) Developed an alternate to Captcha and "I am not a Robot feature used as digital signature in form filling. Engineered a convolutional Neural Network (CNN) model fed with facial features like eyes, nose and lips and voice acoustic data like pitch and frequencv, and attained an accuracv of 77%. Additionally, implemented advertisement recommendation system based on gender detection aiming better marketing.