Analysis of the Impact of Dimensionality Reduction on Deep Learning Performance for the Classification of Microarray and RNA-Seq Gene Expression Data

Duwi Lufita Marfiana; Muhammad Haris

doi:10.34148/teknika.v15i1.1400

Authors

Duwi Lufita Marfiana Computer Science Department, Faculty of Information Technology, Universitas Nusa Mandiri, DKI Jakarta, Indonesia
Muhammad Haris Computer Science Department, Faculty of Information Technology, Universitas Nusa Mandiri, DKI Jakarta, Indonesia

DOI:

https://doi.org/10.34148/teknika.v15i1.1400

Keywords:

Dimension Reduction, Deep Learning, Gene Expression, RNA-Seq, Bioinformatics

Abstract

One of the most popular and widely used methods in various fields, not only in information technology, including bioinformatics. With the increasing complexity of data and problems faced, ML offers solutions for analyzing and interpreting very large and dense data. This study aims to analyze gene expression, which has a crucial role in disease diagnosis, especially cancer, requiring techniques capable of identifying patterns and relationships in complex genetic data. Several techniques that play a role in improving the performance of deep learning models, including PCA reduction techniques, TruncatedSVD, PaCMAP, and TorchDR, successfully simplified data with high dimensionality without eliminating important information structures, thereby speeding up the training process and improving prediction accuracy. DNN was able to achieve a perfect score with accuracy and F1-Score 1 and remained superior to TorchDR with accuracy using test data of 0.9990, PaCMAP 0.979, and TruncatedSVD 0.988. DNN is able to obtain a perfect score with accuracy and F1-Score 1, and remains superior to TorchDR with accuracy using test data of 0.9990, PaCMAP 0.979 and TruncatedSVD 0.988. The dimensionality reduction techniques of PaCMAP and TorchDR provide a more stable data representation, by producing a more separated data distribution and supporting high performance in DNN and MLP models, which makes it the right choice in the dimensionality reduction method of gene expression data.

Downloads

Download data is not yet available.

References

[1] B. Bagiroz, E. Doruk, and O. Yildiz, “Machine learning in bioinformatics: Gene expression and microarray studies,” in Proc. 2020 Medical Technologies Congress (TIPTEKNO), 2020, doi: 10.1109/TIPTEKNO50054.2020.9299285.

[2] N. Tabassum, M. A. S. Kamal, M. A. H. Akhand, and K. Yamada, “Cancer classification from gene expression using ensemble learning with an influential feature selection technique,” BioMedInformatics, vol. 4, no. 2, pp. 1275–1288, 2024, doi: 10.3390/biomedinformatics4020070.

[3] X. Deng and Y. Xu, “Cancer classification using microarray data by DPCAForest,” in Proc. IEEE 31st Int. Conf. Tools with Artificial Intelligence (ICTAI), Portland, OR, USA, 2019, pp. 1081–1087, doi: 10.1109/ICTAI.2019.00151.

[4] H. Chebli, Z. Mashhadieh, M. A. Ali, M. K. Madi, and I. R. Kassem, “Unlocking the potential of DNA microarray for accurate cancer diagnosis with deep learning,” in Proc. 7th Int. Conf. Advances in Biomedical Engineering (ICABME), 2023, doi: 10.1109/ICABME59496.2023.10293017.

[5] H. S. Basavegowda and G. Dagnew, “Deep learning approach for microarray cancer data classification,” CAAI Trans. Intell. Technol., vol. 5, no. 1, pp. 22–33, 2019, doi: 10.1049/trit.2019.0028.

[6] S. Gupta, M. K. Gupta, M. Shabaz, and A. Sharma, “Deep learning techniques for cancer classification using microarray gene expression data,” Frontiers in Physiology, vol. 13, 2022, doi: 10.3389/fphys.2022.952709.

[7] Y. Zeng, Y. Zhang, Z. Xiao, and H. Sui, “A multi-classification deep neural network for cancer type identification from high-dimension, small-sample and imbalanced gene microarray data,” Scientific Reports, vol. 15, no. 1, 2025, doi: 10.1038/s41598-025-89475-2.

[8] B. C. Feltes, E. B. Chandelier, B. I. Grisci, and M. Dorn, “CuMiDa: An extensively curated microarray database for benchmarking and testing of machine learning approaches in cancer research,” J. Comput. Biol., vol. 26, no. 4, pp. 376–386, 2019, doi: 10.1089/cmb.2018.0238.

[9] M. Ilyas, K. M. Aamir, S. Manzoor, and M. Deriche, “Linear programming based computational technique for leukemia classification using gene expression profile,” PLoS ONE, vol. 18, no. 10, p. e0292172, 2023, doi: 10.1371/journal.pone.0292172.

[10] H. Z. Almarzouki, “Deep-learning-based cancer profiles classification using gene expression data profile,” J. Healthcare Eng., vol. 2022, pp. 1–13, 2022, doi: 10.1155/2022/4715998.

[11] S. Sharma, S. Sharma, and A. Athaiya, “Activation functions in neural networks,” Int. J. Eng. Appl. Sci. Technol., vol. 4, no. 12, pp. 310–316, 2020, doi: 10.33564/IJEAST.2020.V04I12.054.

[12] B. Škrlj, S. Džeroski, N. Lavrač, and M. Petković, “Feature importance estimation with self-attention networks,” in Proc. Eur. Conf. Artificial Intelligence (ECAI), 2020, pp. 1491–1498. [Online]. Available: https://dblp.uni-trier.de/db/conf/ecai/ecai2020.html#SkrljDLP20

[13] D. S. R. Polamuri, M. Kumbhkar, and D. A. P. Daniel, Introduction to Deep Learning, 1st ed. India: AGPH Books, 2022, ISBN: 978-93-94339-21-7.

[14] I. Zeger and G. Sisul, “Introduction to deep learning possibilities in communication systems,” in Proc. Int. Symp. ELMAR, vol. 31, 2021, pp. 21–24, doi: 10.1109/ELMAR52657.2021.9550825.

[15] A. A. Alnuaim et al., “Human-computer interaction for recognizing speech emotions using multilayer perceptron classifier,” J. Healthcare Eng., vol. 2022, pp. 1–12, 2022, doi: 10.1155/2022/6005446.