Burkholderia Pseudomallei Infection Detection - Rule Explanation

Array generation approach

The data used for using these two rules were obtained using a proteome array chip to measure antibody responses to a panel of 214 immunoreactive antigens. Only relevant genes are listed in the rule boxes. Sera from melioidosis positive or negative patients in Singapore were generally taken on admission to hospital or from walk-in clinics. Positive samples were taken from patients on admission to hospital and who had a diagnosis of melioidosis confirmed by blood culture.

Simple rule

The simple rule uses three anti-genes (BPSS1512, BPSL1445, and BPSL2697) for the diagnosis. The diagnostic function is a Boolean OR function of them. For diagnosis, you need to input the expression levels of these three antigenes sampled from a patient. The system will tell you if the patient has been infected by Burkholderia Pseudomallei. The estimated prediction accuracy is 91.95%.

Complex rule

The complex rule uses 18 anti-genes for the diagnosis. The diagnostic function is a Boolean OR function of them. The Boolean function is expressed as

BPSL1901 and BPSL2697 or BPSL3341 and BPSL2697 or BPSS1531 and BPSL2697 or BPSL2615 and BPSL2697 or (BPSS2288 or BPSL1317 or BPSS1512 and BPSS0477) or BPSL2017 or BPSL1745 or BPSL2522 or BPSS0530 or BPSL2520 or BPSS2053 or BPSL2298 or BPSS1534

For diagnosis, you need to input the expression levels of these 18 antigenes sampled from a patient. The system will tell you if the patient has been infected by Burkholderia Pseudomallei. The estimated prediction accuracy is 96.55%. Click here to use it

Rule explanation

Each gene expression level is firstly converted to a binary value using the foolowing equation:

The graphs below show the density functions before and after applying the logarithm