I am not really a specialist in FTIR but methods for spectra deconvolution can be found in the literature, for example D.G. Cameron, D.J. Moffat, J. Test. Eval. 12 (1984) 78-85. In addition I found a highly recommended reply by Hatsuo Ishida on ResearchGate of about 10 years ago (see attachment) who stresses the importance of selecting bands with known structural origin for deconvolution.

In analysing starch-protein-water mixtures, you could focus on the starch region at 950-1050 cm-1 and the amide I and II bands at 1550 and 1650 cm-1 (Internet). These regions are well documented and are not likely to show too much interference. For specialised literature on FTIR of starch, see J.J.G. van Soest et al., Carbohydr. Res. 279 (1995) 201-214, and I. Capron et al., Carbohydr. Polym. 68 (2007) 249-259. In order to investigate possible overlap of starch and protein signals you could consider to perform experiments on well-documented starch and protein samples in the dried state, in pure form and in synthetic mixtures, and/or remove one of the two from the mixture by action of glucoamylase or protease and dialysis.

In addition to my previous reply: by searching for 'ftir deconvolution' I found a lot of hits including video's. See the attachment.

I am extremely sorry Sir for the late reply due to health issues, thank you so much Sir for your kind assistance and guidance in this, I will look into these and let you know where I am sticking behind in this.