Back

Previously unrecognized non-reproducible antibody-antigen interactions and their implications for diagnosis of viral infections including COVID-19

Pan, J.; Yang, L.; Deng, Y.; Sun, B.; Zhang, L.; Wu, W.; Li, J.; Cheng, H.; Li, Y.; Xu, W.; Yang, J.; Sun, Y.; Fei, H.; Xue, Q.; Zhou, Y.; Wang, H.; Zheng, P.; Chen, H.; Zhu, F. C.; Peng, D.; Gu, D.; Han, J.; Qiu, J.; Ma, H.

2021-07-20 immunology
10.1101/2021.07.20.453011 bioRxiv
Show abstract

Antibody-antigen (Ab-Ag) interactions are canonically described by a model which exclusively accommodates non-interaction (0) or reproducible-interaction (RI) states, yet this model is inadequate to explain often-encountered non-reproducible signals. Here, by monitoring diverse experimental systems and confirmed COVID-19 clinical sera using a peptide microarray, we observed that non-specific interactions (NSI) comprise a substantial proportion of non-reproducible antibody-based results. This enabled our discovery and capacity to reliably identify non-reproducible Ab-Ag interactions (NRI), as well as our development of a powerful explanatory model ("0-RI-NRI-Hook four-state model") that is [mAb]-dependent, regardless of specificity, which ultimately shows that both NSI and NRI are not predictable yet certain-to-happen. In experiments using seven FDA-approved mAb drugs, we demonstrated the use of NSI counts in predicting epitope type. Beyond challenging the centrality of Ab-Ag interaction specificity data in serology and immunology, our discoveries also facilitated the rapid development of a serological test with uniquely informative COVID-19 diagnosis performance.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.