U Texas say some antibody tests for Covid-19 detect the wrong antibody to predict future immunity

https://medicalxpress.com/news/2020-08-frequently-serology-antibodies-reinfection-covid-.html

Sample quote

Results showed that 3% of healthy and non-COVID-19 samples collected during the pandemic in Houston were positive for the N-protein antibody, but only 1.6% of those had the S-RBD antibody. Of samples with the S-RBD antibody, 86% had neutralizing capacity—meaning they could prevent reinfection of COVID-19, but only 74% of samples with N-protein had neutralizing capacity. When positive for both, 96.5% exhibited neutralizing capacity.

"These findings suggest that detection of N-protein binding antibodies does not always correlate with presence of S-RBD neutralizing antibodies, and that the presence of the S-RBD antibody is the best indicator of any potential protection against reinfection," said senior author Raghu Kalluri, M.D., Ph.D., professor and chair of Cancer Biology. "We caution against the extensive use of N-protein based serology testing for determination of potential COVID-19 immunity, and we believe that accurate and reliable S-RBD serological testing is needed to carefully identify individuals with neutralizing antibodies in order to help advance recovery efforts around the globe."

End quote

T-cells are not detected by antibody tests either