Use of multivariate analysis to assess treatment approaches

A method to predict changes in adaptive behavior comprising:

• performing Fisher discriminant analysis on one or more data sets of typically developing (TD) individuals and one or more data sets of autism spectrum disorder (ASD) individuals to produce a model that classifies TD individuals from ASD individuals, wherein the TD data sets include folate-dependent one-carbon metabolism (FOCM) and transsulfuration (TS) pathway metabolic profile data for a plurality of TD individuals and ASD data sets include pre-treatment folate-dependent one-carbon metabolism (FOCM) and transsulfuration (TS) pathway metabolic profile data and post-treatment folate-dependent one-carbon metabolism (FOCM) and transsulfuration (TS) pathway metabolic profile data for patients receiving one or more ASD treatments;
• calculating, for the plurality of patients in the ASD data sets, a first probability that the model will classify a data set patient as TD pre-treatment and a second probability that the model will classify the data set patient as TD post-treatment;
• identifying an increase from the first probability to the second probability to quantify a positive treatment effect by at least one of the ASD treatments on metabolic profiles of the plurality of patients;
• administering a particular ASD treatment associated with a quantified positive treatment effect to a target patient;
• performing a treatment effect regression analysis on changes in adaptive behavior scores and metabolic profiles of the plurality of patients;
• and
• quantifying a predicted adaptive behavior score change for the target patient from a measured metabolic profile change in the target patient,
• wherein the particular ASD treatment includes MeCbl+LDFA, BH4, HDFA, or combinations thereof.