Page 7 - IJB-7-1

An, et al.

Figure 2. A vision for future bioprinting.

recently reviewed in Zhang et al. , Schwab et al. . These huge and diverse, it could be all types of diagnostic images
[12]
[11]
mathematical models are useful for construction of virtual stored in hospital databases, all types of experimental
bioprinting process. ML models have also been used in data in worldwide laboratories and research centers, all
in vitro study for identification of cell signature genes out the “omics” databases already established in past years,
of complex gene expression profiles among different cell or simply the vast scientific literature. Standard and open-
groups [13,14] . Another in vitro example is virtual histological access databases with meaningful and valuable training
staining, which bypasses the lengthy and laborious process datasets specifically targeting for bioprinting must be
for tissue preparation. Researchers used deep learning to created from Big Data curation. A recent example is
transform autofluorescence images of tissue into images the construction of a web-based nanomaterial database
equivalent to histologically stained tissue , and achieved through Big Data curation, which contains 705 unique
[15]
blending of multiple stains by assigning each stain at the nanomaterials, and the annotation of nanostructures
pixel level [16,17] . Furthermore, mathematical models and generates 2142 nanodescriptors for modeling and ML, but
[20]
ML models, which help us understand the complexities of more importantly, the database is publicly available .
biological systems and extract new biological knowledge Another example is geoscience databases, which is large
[21]
from complex experimental datasets, are expected to and ideal for ML and automated geoscience analysis .
bring tissue engineering much closer to clinical reality . In fact, numerous experimental data and various materials
[18]
Collectively, the above evidence of virtual experiments directly related to bioprinting have been generated over
in either processing or post-processing of bioprinting past years, making bioprinting a potentially a data-
suggests that we will see more in silico experiments with driven research, but so far there is limited database
ML in bioprinting in future. created specifically for bioprinting. In future, we hope
However, ML cannot be performed without Big to see more developments in this area, in line with the
Data about modern clinical imaging of organs, histology, development of databases for 3D printing. Perhaps it may
immunohistochemistry, biomechanical properties of be even possible to predict new bioprinting discoveries
tissue and organs, molecular profiles of cell, tissue, and by exploiting the current literature alone without relying
[22]
organs (genomics and proteomics) function and so on. Big on experts’ opinions .
Data can be structured, unstructured or semi-structured 4. Digital twin of human organ
and it is much more than traditional databases . For ML
[19]
purposes, the first step is collection of Big Data or Big On the other hand, design process in 3D bioprinting must
Data curation. The sources of big data for bioprinting are be organized around the concept of digital twins of organs

International Journal of Bioprinting (2021)–Volume 7, Issue 1 3