Method and system for risk assessment of Polycystic Ovarian Syndrome (PCOS)
This disclosure relates generally to and, more particularly, to assessment of PCOS. Polycystic ovarian syndrome (PCOS) is a hormonal disorder common among women of reproductive age that causes infertility and affects overall health of the woman. As PCOS is common and curable cause of infertility, an efficient early screening to assess a potential risk of PCOS can ensure early treatment. The current state-of-the-art techniques include diagnostic, screening solutions, imaging techniques which are invasive, complex, expensive. The disclosure is a supervised machine learning algorithm on the samples of individuals to arrive at a panel of biological features/indicators/markers/signatures that can accurately stratify/classify/group individuals into ‘PCOS’ and ‘healthy’ based upon the differences in the composition of the gut/oral microbial communities.
1 . A method for risk assessment of poly-cystic ovarian syndrome (PCOS) disorder of a subject, comprising the steps of:
collecting a saliva sample and a stool sample of the subject whose risk of PCOS disorder is to be assessed;
extracting microbial deoxyribonucleic acid (DNA) sequences from each of the saliva sample and the stool sample, individually;
determining a quantitative abundance of: (i) each of a plurality of predetermined microbes associated with the saliva sample and (ii) each of a plurality of predetermined microbes associated with the stool sample, individually, from respective extracted DNA sequences, using a first set of probes and a second set of probes specific to each of the plurality of predetermined microbes associated with the saliva sample and the stool sample respectively, through a multiplex quantitative Polymerase Chain Reaction (qPCR) technique;
collating, via one or more hardware processors, the quantitative abundance of: (i) each of the plurality of predetermined microbes associated with the saliva sample and (ii) each of the plurality of predetermined microbes associated with the stool sample, to obtain a hybrid abundance matrix;
determining, via the one or more hardware processors, a model score based on the hybrid abundance matrix, using a pre-determined machine learning (ML) model;
performing, via the one or more hardware processors, risk assessment of the subject, based on the model score and a predefined threshold value; and
providing a designed personalized microbial concoction to the subject to enable amelioration of PCOS based on the risk assessment of the subject, wherein the designed personalized microbial concoction for the subject consisting of a specific composition of microbes employable for the subject assessed as having poly-cystic ovarian syndrome (PCOS) disorder for ameliorating the PCOS disorder.
2 . The method of claim 1 , wherein:
(i) the plurality of predetermined microbes associated with the saliva sample comprises: Phocaeicola, Simonsiella, Massiliprevotella, Streptobacillus, Rothia , and Slackia ; and
(ii) the plurality of predetermined microbes associated with the stool sample comprises: Oscillibacter, Cuneatibacter, Faecalibacterium, Pseudoflavonifractor, Gordonibacter, Neisseria, Fusobacterium, Lactococcus , and Lactobacillus.
3 . The method of claim 1 , wherein the first set of probes specific to each of the plurality of predetermined microbes associated with the saliva sample are utilized in a first multiplex qPCR run, and a second multiplex qPCR run, to determine the quantitative abundance of each of the plurality of predetermined microbes associated with the saliva sample, and wherein:
(i) the plurality of predetermined microbes, the quantitative abundance of which are being determined through the first multiplex qPCR run are: Phocaeicola, Simonsiella, Massiliprevotella and Streptobacillus ; and
(ii) the plurality of predetermined microbes, the quantitative abundance of which are being determined through the second multiplex qPCR run are: Phocaeicola, Simonsiella, Rothia , and Slackia.
4 . The method of claim 1 , wherein the second set of probes specific to each of the plurality of predetermined microbes associated with the stool sample are utilized in a third multiplex qPCR run, fourth multiplex qPCR run, and a fifth multiplex qPCR run, to determine the quantitative abundance of each of the plurality of predetermined microbes associated with the stool sample, and wherein:
(i) the plurality of predetermined microbes, the quantitative abundance of which are being determined through the third multiplex qPCR run are: Oscillibacter, Cuneatibacter, Pseudoflavonifractor , and Gordonibacter , and
(ii) the plurality of predetermined microbes, the quantitative abundance of which are being determined through the fourth multiplex qPCR run are: Oscillibacter, Cuneatibacter, Faecalibacterium and Neisseria ; and
(iii) the plurality of predetermined microbes, the quantitative abundance of which are being determined through the fifth multiplex qPCR run are: Fusobacterium, Lactococcus, Faecalibacterium , and Lactobacilus.
5 . The method of claim 1 , wherein the pre-determined machine learning (ML) model is an ensemble ML model that is built using a microbial abundance data corresponding to a plurality of training saliva samples and a plurality of training stool samples.
6 . The method of claim 1 , wherein the plurality of predetermined microbes associated with the saliva sample and the plurality of predetermined microbes associated with the stool sample are features of the pre-determined machine learning (ML) model.
7 . The method of claim 3 , wherein one or more predetermined microbes out of the plurality of predetermined microbes associated with the saliva sample, are common to the first multiplex qPCR run and the second multiplex qPCR run for determining the quantitative abundance, and wherein the one or more predetermined microbes that are common to the first multiplex qPCR run and the second multiplex qPCR run are determined based on (i) a median abundance of each of the plurality of predetermined microbes obtained from the plurality of training saliva samples, (ii) a frequency of occurrence of each of the plurality of predetermined microbes constituting the ensemble ML model.
8 . The method of claim 4 , wherein one or more predetermined microbes out of the plurality of predetermined microbes associated with the stool sample are common to the third multiplex qPCR run and the fourth multiplex qPCR run for determining the quantitative abundance, and wherein the one or more predetermined microbes that are common to the third multiplex qPCR run and the fourth multiplex qPCR run are determined based on (i) a median abundance of each of the plurality of predetermined microbes obtained from the plurality of training stool samples, (ii) a frequency of occurrence of each of the plurality of predetermined microbes constituting the ensemble ML model.
9 . The method of claim 4 , wherein one or more predetermined microbes out of the plurality of predetermined microbes associated with the stool sample are common to the fourth multiplex qPCR run and the fifth multiplex qPCR run for determining the quantitative abundance, and wherein the one or more predetermined microbes that are common to the fourth multiplex qPCR run and the fifth multiplex qPCR run are determined based on (i) a median abundance of each of the plurality of predetermined microbes obtained from the plurality of training stool samples, (ii) a frequency of occurrence of each of the plurality of predetermined microbes constituting the ensemble ML model.
10 . The method of claim 6 , wherein designing the personalized microbial concoction comprises designing one or a combination of probiotic and antibiotic microbial candidates based on the features constituting the ML model, and wherein designing of the one or the combination of probiotic and antibiotic candidates is performed by mapping the features constituting the ML model to complete set of microbes or a pre-defined subset of the microbes, wherein the features comprise microbial taxa.
11 . The method of claim 10 , wherein designing of the one or the combination of probiotic and antibiotic microbial candidates comprises:
1) Computing pair-wise correlations between abundances of the microbial taxa constituting the ML model and the abundances corresponding to the complete set of microbial taxa represented by microbial markers computed individually from (a) the subset of training samples corresponding to a healthy class (A) that is a class of samples that were taken from patients diagnosed to be free of PCOS, and (b) a diseased class (B) that is a class of samples that were taken from PCOS affected subjects, wherein both the samples belonging to the healthy class (A) and the diseased class (B) are used as training data for generating the ML model;
2) positive and negative interactions between the microbial taxa constituting the ML model and all other taxa in the healthy class and the diseased class of the training samples are deduced using critical correlation (r) value as the cut-off such that inter-taxa correlation index values greater than +r value are affiliated as ‘positive interactions’, while those less than-r value are affiliated as ‘negative interactions’;
3) repeating steps 1 and 2 for a predefined numbers of times and considering only those interactions as relevant that appear in at least 70% of iterations with the critical correlation (r) value cut-off and retained the relevant interactions;
4) using a set of predefined rules to select relevant microbial candidates for designing the probiotic and/or antibiotic microbial candidates based on the retained model taxa interactions.
12 . The method of claim 11 , wherein the microbial markers in the ML model which are higher in the diseased class are suitable antibiotic targets for which existing antibiotics to be used or designed from the microbial taxa which are having the negative interactions with target marker microbe, and the probiotics are targeted at the microbial markers found to be abundant in case of healthy class subjects in the training model, where existing probiotics to be used or designed from the microbes that are positively interacting with the target marker microbe.