Biology subjects

Hari, S. N.

Publications and source records attributed to Hari, S. N..

3 recordsLinked to original sources

bioRxiv2025

Exploring non-invasive sexing of early chick embryos in intact eggs using Laser Speckle Contrast Imaging (LSCI) and Deep Neural Network (DNN)

The ability to image blood flow in early-stage avian embryos has significant applications in developmental biology, drug and vaccine testing, as well as determining sex differentiation. In this project, we used our recently developed laser speckle contrast imaging (LSCI) system to non-invasively image extraembryonic blood vessels and used these images to attempt early sex identification of chick embryos. Specifically, we captured images of blood vessels from 1,251 living chicken embryos between day three and day four of incubation. We then applied deep neural network (DNN) models to evaluate whether it is possible to differentiate sex based on vascular patterns. Using ResNetBiT and YOLOv5 models, our results indicate that sex differentiation from extraembryonic blood vessel images was not achievable with sufficiently high accuracy or statistical significance for practical use. Specifically, ResNetBiT had a five-fold cross-validated average accuracy of 59%{+/-}5% (fold-wise p-value, p [≤] 0.3) at day 3 and 61%{+/-}3% (fold-wise, p [≤] 0.04) at day 4. YOLOv5 had a five-fold cross-validated average accuracy of 55%{+/-}3% (fold-wise, p [≤] 0.3) at day 3 and 53%{+/-}3% (fold-wise, p [≤] 0.5) at day 4. Our findings suggest that using vascular pattern imaging alone is inconclusive for reliable early sex identification in chicken embryos.

bioengineering↗

bioRxiv2023

Spatially aware deep learning reveals tumor heterogeneity patterns that encode distinct kidney cancer states

Background Abstract Background Results Discussion Methods Data Availability Code Availability Author Contributions Competing Interests Extended Data References Renal cell carcinoma (RCC) is among the 10 most common cancers worldwide and is comprised of several histological subtypes1. The clear cell histological subtype (ccRCC) is the most common form of RCC and accounts for the vast majority (75-80%) of metastatic cases1. In addition to highly recurrent mutations in hypoxia (VHL) and chromatin regulator genes (e.g. PBRM1, BAP1, SETD2), ccRCC exhibits extensive genomic intratumoral heterogeneity (ITH)2, which wa ...

cancer biology↗

bioRxiv2021

Examining Batch Effect in Histopathology as a Distributionally Robust Optimization Problem

Computer vision (CV) approaches applied to digital pathology have informed biological discovery and development of tools to help inform clinical decision-making. However, batch effects in the images have the potential to introduce spurious confounders and represent a major challenge to effective analysis and interpretation of these data. Standard methods to circumvent learning such confounders include (i) application of image augmentation techniques and (ii) examination of the learning process by evaluating through external validation (e.g., unseen data coming from a comparable dataset collected at another hospital). Here, we show that the source site of a histopathology slide can be learned from the image using CV algorithms in spite of image augmentation, and we explore these source site predictions using interpretability tools. A CV model trained using Empirical Risk Minimization (ERM) risks learning this source-site signal as a spurious correlate in the weak-label regime, which we abate by using a training method with abstention. We find that a patch based classifier trained using abstention outperformed a model trained using ERM by 9.9, 10 and 19.4% F1 in the binary classification tasks of identifying tumor versus normal tissue in lung adenocarcinoma, Gleason score in prostate adenocarcinoma, and tumor tissue grade in clear cell renal cell carcinoma, respectively, at the expense of up to 80% coverage (defined as the percent of tiles not abstained on by the model). Further, by examining the areas abstained by the model, we find that the model trained using abstention is more robust to heterogeneity, artifacts and spurious correlates in the tissue. Thus, a method trained with abstention may offer novel insights into relevant areas of the tissue contributing to a particular phenotype. Together, we suggest using data augmentation methods that help mitigate a digital pathology models reliance on potentially spurious visual features, as well as selecting models that can identify features truly relevant for translational discovery and clinical decision support.

cancer biology↗