From that same paper, please note that only 2/3 of the tested samples contained virus detectable by RT-PCR--which only detects nucleic acids, and cannot differentiate between viable and dead virus. Out of those 8 samples positive by RT-PCR, virus could only be cultured out of one sample. That means that only 8.33% of the patients sampled had viable virus in their saliva.
In any case, you should not kiss or hug Ebola patients without full PPE.
Informed consent was obtained from the patient or guardian. A convenience sample of various clinical specimens, primarily bodily fluids, was obtained from patients with laboratory-confirmed EHF
The color and absence or presence of blood was noted for each sample. All specimens were placed into sterile cryovials and stored at ambient temperature (∼25°C–30°C) in the isolation ward for the rest of the day (typically ⩽6 h) before being stored in liquid nitrogen at the field laboratory established for the outbreak.
There was a significant discrepancy between the results of virus culture and RT-PCR testing in our study, with many more frequent positive results from RT-PCR. Possible explanations for this finding include virus degradation from breaks in the cold chain during sample collection, storage, and shipping; the greater sensitivity of RT-PCR relative to culture; and, in the case of the saliva specimens, possible virus inactivation by salivary enzymes. The less-than-ideal storage conditions of the specimens in the isolation ward immediately after acquisition and the fact that even the nasal blood from 1 patient was culture negative suggest that some virus degradation indeed occurred.
http://www.ncbi.nlm.nih.gov/pubmed/16652308
METHODS: Serum and oral fluid specimens were obtained from 24 patients with suspected Ebola and 10 healthy control subjects. Specimens were analyzed for immunoglobulin G antibodies by enzyme-linked immunosorbent assay (ELISA) and for Ebola virus by antigen detection ELISA and reverse-transcriptase polymerase chain reaction (RT-PCR). Oral fluid specimens were collected with a commercially available collection device. RESULTS: We failed to detect antibodies against Ebola in the oral fluid specimens obtained from patients whose serum samples were seropositive. All patients with positive serum RT-PCR results also had positive results for their oral fluid specimens.
which shows that pcr detected parts of either active or inactive viral particles in the saliva