An indexed modeling and experimental strategy for biosignatures of pathogen and host

In information theory, a signature is characterized by the information content as well as noise statistics of the communication channel. Biosignatures have analogous properties. A biosignature can be associated with a particular attribute of a pathogen or a host. However, the signature may be lost in backgrounds of similar or even identical signals from other sources. In this paper, we highlight statistical and signal processing challenges associated with identifying good biosignatures for pathogens in host and other environments. In some cases, it may be possible to identify useful signatures of pathogens through indirect but amplified signals from the host. Discovery of these signatures requires new approaches to modeling and data interpretation. Finally, an understanding of pathways (or gene networks) that are modified through host–pathogen interactions will result in better detectors as well as opportunities in vaccines and therapeutics.