International Journal of Food Sciences and Nutrition,

February 2012; 63(1): 5–15

Enhancement of nutritional value of finger millet-based food

(Indian dosa) by co-fermentation with horse gram flour

BRUNTHA DEVI PALANISAMY1, VIJAYABHARATHI RAJENDRAN1, SATHYABAMA

SATHYASEELAN1, RAJEEV BHAT2, & BRINDHA PRIYADARISINI VENKATESAN1
1
Department of Microbial Biotechnology, Bharathiar University, Coimbatore 641 046, India, and 2Food Technology Division,
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13

School of Industrial Technology, Universiti Sains Malaysia, Penang 11800, Malaysia

Abstract
Co-fermentation of finger millet with horse gram was carried out to produce inexpensive protein-rich food (dosa-an Indian
breakfast food). Natural fermentation of finger millet –horse gram flour blend in different proportions (2:1, 3:1, 4:1 and 5:1) was
performed for 24 h. Biochemical analysis showed reasonable drop in pH (6.6 – 4.2) and starch content (25.52%) with
considerable augment in titratable acidity (0.168 – 1.046%), soluble proteins (1.1-fold) and free amino acids (2.6-fold) at 16 h.
Lactic acid bacteria dominated yeast counts throughout the fermentation accompanied by a decrease in total soluble and
reducing sugars. Total essential amino acids increased 1.1-fold at 16-h fermentation with protein containing 48.68% of essential
amino acids over total amino acids. Lysine increased from 5.87 to 6.73 g of amino acid/100 g of total amino acids. Dosa, prepared
from 16-h fermented batter, showed better sensory attributes for 4:1 ratio. The formulated new product might be used to
For personal use only.

overcome the protein – energy malnutrition problems.

Keywords: finger millet, horse gram, co-fermentation, traditional fermented food, dosa

Introduction
Cereals such as wheat, rice, maize, sorghum and minor
millets like pearl millet and finger millet are important
sources of dietary proteins, carbohydrates, vitamins,
minerals and fibre for the people residing in Asian
and African sub-continents (Blandino et al. 2003).
However, they offer several challenges from the
nutritional point of view, especially swelling of their
starch on cooking and limited bioavailability of amino
acids and minerals due to the presence of anti-
nutritional compounds that tend to reduce bioavail-
ability by 5 –15% (Nout 2009). Several processing
methods such as cooking, germination, milling and
fermentation are routinely employed to improve the
nutritional properties of cereals. Among these,
fermentation has proved to be the best processing
method (Blandino et al. 2003). Fermentation not only
improves the sensory characteristics of a product, but
also eliminates certain undesirable constituents,
makes the nutrients more accessible while preserving
and enhances the levels of many bioactive compounds
(Hubert et al. 2008). Co-fermentation of cereals
with legumes has often been proposed to produce
inexpensive protein-rich foods for improving the
macro- and micronutrient balance of cereal-based
fermented food products. Generally, cereals are
deficient in lysine, but are rich in cysteine and
methionine, whereas legumes are rich in lysine, but
deficient in sulphur-containing amino acids. There-
fore, by combining cereals with legumes, the overall
protein quality can be improved. The most popular
traditionally consumed cereal –legume-based fermen-
ted foods in India include adai, dhokla, dosa, idli and
vadai (Chavan and Kadam 1989).
Finger millet (Eleusine coracana) constitutes the
staple food of people belonging to the low socio-
economic group. It has a carbohydrate content of

Correspondence: P. Bruntha Devi, Department of Microbial Biotechnology, Bharathiar University, Coimbatore 641 046, India.
E-mail: brundha.biotech@gmail.com

ISSN 0963-7486 print/ISSN 1465-3478 online q 2011 Informa UK, Ltd.

DOI: 10.3109/09637486.2011.591367
 6 B. D. Palanisamy et al.
81.5%, protein 9.8%, crude fibre 4.3% and mineral
2.7%. Nowadays, it is recognized as neutraceuticals
due to its high content of polyphenols (0.3 –3%) and
dietary fibre (19%), which offer health benefits such as
hypoglycaemic, hypocholesterolaemic, anti-microbial
and anti-oxidant properties. However, maximum
utilization of the nutrient potential of the millet is
limited by the presence of phytates (0.48%), tannins
(0.61%) and enzyme inhibitors (Ravindran 1992).
While phytates bind essential minerals and proteins,
tannins complex with proteins and enzyme inhibitors
reduce digestibility (Reddy and Pierson 1994) and
fermentation enhances the biological value, net
protein utilization, thiamin, riboflavin and niacin
contents in finger millet (Aliya and Geervani 1981).
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
Auto-fermentation of finger millet significantly
reduces the anti-nutritional components with simul-
taneous increase in mineral extractability, in vitro
protein and starch digestibility. Traditionally, it is
consumed in the form of thick porridge (mudde or
dumpling), thin fermented porridge (ambali), fried or
baked pancake (roti, dosa) and beverages (chang/jnard)
(Malleshi and Hadimani 1993), and most of these are
prepared by fermentation. Since it is consumed whole,
fibre, minerals, phenolics and vitamins present in the
outer layer of the grain or the seed coat are retained,
For personal use only.
which offer health benefits such as prevention of many
chronic diseases (Antony et al. 1996).
However, studies on the co-fermentation of finger
millet with legumes and their fermented food products
are scarce. Hence, our present study mainly focuses
and incorporates co-fermentation of finger millet with
a cost-effective protein source such as horse gram
(Macrotyloma uniflorum L.), which is also known as
poor man’s pulse in Southern India. It is an excellent
source of proteins (17.9 – 25.3%), carbohydrates
(51.9 –50.9%), essential amino acids, energy, low
content of lipid (0.58 – 2.06%), iron and molybdenum
(Bravo et al. 1999). It possesses hypoglycaemic and
hypolipidaemic activity. Traditionally, it is consumed
either as a whole seed, sprouts or cooked with rice,
sorghum, pearl millet by soaking or dry heating the
seeds. The use of dry seeds as human food in large
population is limited due to its poor cooking quality,
the presence of high levels of enzyme inhibitors and
haemagglutinin activities (Ray 1969). Utilization can
be made significant by implementation of economi-
cally viable diverse processing strategies. Although
soaking, germination, dry heating, popping and
microwave cooking methods have been adopted for
processing horse gram, the extent of fermentation to
which they enhance the nutritional values was not
carried out till date. Fermentation may improve the
texture and palatability, and inactivates the anti-
nutritional factors in legume-based products.
Usually, idli and dosa products are based on rice-
black gram blends; this study reports the alternative
of this with finger millet – horse gram. This study has
been carried out to determine the biochemical,
nutritional and microbial population changes during
co-fermentation of finger millet and horse gram flour
along with the evaluation of sensory attributes of
fermented food product (Indian dosa) prepared from
the blended batter. This is expected to provide more
insight on the co-fermentation of cereals with legumes,
and improves the overall nutritional status of
fermented foods. Additionally, it is envisaged that the
development of a new product might prove to be
useful to overcome the protein-energy malnutrition
problems in developing countries.
Methods
Materials
Finger millet Co. (Ra)-14 brown variety was pur-
chased in bulk from the Millet Breeding Station, Tamil
Nadu Agricultural University, Coimbatore, India, and
horse gram was purchased from the local market as a
single batch. The grains were cleaned, washed, dried
and stored in airtight containers. All the chemicals and
organic solvents used in this study were of analytical
grade (Merck, Mumbai, India). The growth media for
microbial analysis were procured from HiMedia
laboratories, Mumbai, India.
Proximate analysis of the raw materials
The proximal analysis of finger millet and horse gram
was determined. The moisture content was measured
using Moisture Analyser MA30 (Sartorius AG,
Goettingen, Germany), crude protein, crude fibre,
crude fat and total ash content by AOAC method
(2000), and carbohydrate content was determined as
the weight difference (Total carbohydrate ¼ 100 2
(Moisture þ crude protein þ crude fibre þ lipid þ
ash). The energy value was determined by multiplying
the percentage of crude protein, crude fat and
carbohydrate by factors 4, 9 and 4, respectively, and
the estimation was recorded as kcal/100 g (Gopalan
et al. 2000).
Preparation of batter
The whole finger millet grains were milled in a flour
miller (100 mesh size). Horse gram seeds were
dehulled by soaking in water for 2 h followed by sun
drying (after draining water). The dried seeds were
dehulled in a traditional dehuller made of stone.
Winnowing was done to separate the cotyledons from
hulls and then milled in the flour miller. Further, the
flour was mixed in different proportions viz. (A) 2:1,
(B) 3:1, (C) 4:1 and (D) 5:1w/w (Finger millet: Horse
gram), and batter (consistency of dosa batter) was
prepared by mixing blended flour (100 g) of each with
120 ml of water and 0.9% w/w salt in a 500 ml beaker.
The beakers were covered with perforated aluminium
foil and fermented at 308C for 24 h. For stimulating
 Nutritional value of finger millet based food (Indian dosa) by co-fermentation with horse gram flour
household conditions of fermentation, neither the
containers nor the grains were sterilized. Aliquots of
the fermenting batter for analyses were drawn up to
24 h with 4 h intervals under aseptic conditions and
stored at 2 708C until analyses. Fresh batter samples
were used for the microbial analysis, pH and titratable
acidity. The fermentations were carried out in
triplicates.
Analysis
Microbial analysis. The microbial analysis of blended
batters was performed by determining the colony
counts of aerobic mesophilic bacteria, lactic acid
bacteria (LAB) and yeasts by pour plate technique
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
on plate count agar, lactic agar and yeast glucose
chloramphenicol agar, respectively. One gram of fresh
fermented batter was homogenized with 9 ml of sterile
saline solution to obtain a uniform suspension and
plated in duplicates after serial dilution. Plate count
agar plates were incubated at 378C for 48 h, lactic
agar plates at 308C for 48 h and yeast glucose
chloramphenicol agar plates at 308C for 3 days.
The colonies that appeared after the incubation time
were counted as colony-forming units (cfu) per gram of
dried batter. Characteristics of the colony were
For personal use only.
observed and representative single colonies were
isolated, subcultured and stored at 48C for further
analysis.
pH and titratable acidity
The pH of the different sets of batter was measured
directly at specific time intervals (4 h) by a digital pH
meter (Elico LI617, India).
Titratable acidity of fermented batter was deter-
mined by titrating 5 g of fermented batter in 50 ml of
distilled water against freshly prepared 0.1 N NaOH,
using 1% phenolphthalein as an indicator. The values
were expressed in g lactic acid equivalents/100 g of
fermented batter.
Determination of total soluble, reducing and non-reducing
sugars
About 100 mg of the fermented batter was extracted
four times with 20 ml of 80% hot ethanol, pooled
extracts were centrifuged at 10,000 rpm for 20 min
and the supernatant was collected. The supernatant
was concentrated, and the volume was made up to
10 ml. Hundred microlitres of the aliquot was assayed
for total soluble sugars by phenol– sulphuric acid
method (Dubois et al. 1956), reducing sugars by
Somogyi method (Somogyi 1952) and non-reducing
sugars by calculating the difference between total
soluble and reducing sugars. Glucose was used as a
standard, and values were expressed in g/100 g of
dried batter.
7
Determination of total starch content
The total starch was determined from alcohol
insoluble residue of fermented batter based on the
method described by McCready et al. (1990). The
residue was solubilized with 52% perchloric acid to
estimate the sugar content by phenol–sulphuric acid
method (Dubois et al. 1956), and values were
multiplied by a factor of 0.9 to arrive at the starch
content. The percentage decrease in starch content
was calculated by the difference between initial and
final content at different time intervals.
Determination of soluble proteins
Soluble proteins were determined by centrifuging the
fermented slurry at 12,000 rpm for 15 min, and the
supernatant was estimated for protein content by
Folin – Ciocalteu reagent as per Lowry’s method
(Lowry et al. 1951). Bovine serum albumin was used
as a standard, and the values were expressed in mg
protein/100 g of dried batter.
Determination of total free amino acids and total amino
acids composition
Free amino acids were determined spectrophotome-
trically by reaction with ninhydrin in the ethanolic
extract prepared as above (Magne and Larher 1992).
A suitable standard curve was prepared with leucine,
and values were expressed in g leucine equiva-
lents/100 g of dried batter.
Individual amino acid content in the fermented
batter was analysed as per the protocol described
by Mbithi-Mwikya et al. (2000) and Onyango et al.
(2004) with some modification. Samples (100 mg)
were hydrolyzed by 6 N HCL in sealed ampoules in an
oven at 1108C for 23 h. The hydrolysate was rapidly
cooled in an ice-water bath and then made up to 1 L
and filtered. An aliquot of the filtrate was evaporated
to dryness at 408C in a rotavapour system. The residue
was redissolved in 0.2 M sodium citrate buffer (pH
2.2), kept in a shaking water bath for 1 h and filtered
through nylon syringe filter (0.2 mM). The filtrate was
then injected into the automated amino acid analyser
(Model 4151, Pharmacia LKB, Alpha Plus, England).
The analyser is a single column ion-exchange
chromatograph, which uses five lithium acetate buffer
solutions of increasing pH and ionic strength, a
regenerating agent and an ion-exchange column.
Separated amino acids were detected by the ninhydrin
colour reaction and photometric detection at 570 nm
for a-amino acids and at 440 nm for the imino acids
in a post-column derivatization step.
Batter volume
Twenty millilitres aliquot of batter from each
proportion was placed in 100 ml measuring cylinders,
 8 B. D. Palanisamy et al.

Table I. Proximate composition of finger millet and horse gram p # 0.05 (Duncan 1955). This analysis was done using
seeds (g/100 g). the SPSS 13.0 for Windows (2000) computer software.
Nutrients* Finger millet Horse gram

Moisture 8.46 ^ 0.05 8.95 ^ 0.03 Results

Crude protein 7.2 ^ 0.07 22.3 ^ 1.30
Total ash 2.93 ^ 0.03 3.79 ^ 0.01 Proximate composition
Crude fibre 2.29 ^ 0.02 1.88 ^ 1.83
Fat 1.77 ^ 0.12 0.67 ^ 0.12 The proximate composition of raw materials (finger
Carbohydrate 77.35 ^ 0.32 62.41 ^ 1.52 millets and horse gram) used in this work is given
Energy (kcal/100 g) 375.13 ^ 1.23 304.03 ^ 1.16 in Table I.
* The above values are mean ^ SD of triplicate values.

Microbial population. The microbial analysis of all the

and rise in the volume during fermentation was noted four batter samples revealed the presence of both
at specific intervals (4 h) (Steinkraus et al. 1967). The bacteria and yeasts throughout the fermentation
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13

increased volume was measured in millilitre, and
results were expressed in % increase of batter volume.
Sensory analysis
Sensory evaluation of the dosa (pancake) prepared
from each blend of 16 h fermented batter was carried
out to determine their organoleptic characteristics.
Dosa was prepared by cooking the thin spread batter
(1 – 5 mm thickness) on a flat heated plate for 2 –3 min
smeared with little cooking oil. Each product was
For personal use only.
process (Table II). The total aerobic bacterial count
in the four different batter combinations increased
from 0 to 16 h, and thereafter counts stabilized with a
slight decrease up to 24 h. In A (2:1) and C (4:1)
samples, no yeast population was detected till 4 and
0 h, respectively, whereas in B (3:1) and D (5:1), yeast
count increased from 0 to 20 h. The lactic acid
bacterial count increased steadily from 0 to 24 h in all
the blended batter samples with counts of 6.04 –
7.99 log cfu/g. Initially, LAB constituted 10, 27.8, 25
and 12.5% and gradually increased to 67.9, 55.2, 50.9

coded with a number, and was served to the panellists.
A panel of 10 pre-trained judges comprising research
scholars, postgraduate students and technical assist-
ants of the laboratory was constituted to evaluate the
products overall acceptability. To determine the
acceptability of the samples, a seven-point hedonic
rating scale, 7: excellent, 6: very good, 5: good, 4:
satisfactory, 3: fair, 2: poor, 1: very poor, was used.
Statistical analysis
All the experiments were carried out in triplicates, and
the results were represented as mean ^ SD. Data were
assessed by analysis of variance and the significant
differences were found, the mean was separated by
Duncan’s multiple range tests with a probability of
and 63.6% in A, B, C and D, respectively.
Biochemical changes occurring during fermentation
pH and titratable acidity. The fermentation of finger
millet – horse gram flour blend resulted in considerable
drop in pH with a corresponding increase in the
titratable acidity in all four combinations. The initial
of pH 6.6 dropped significantly ( p , 0.05) to 4.2 at
the end of fermentation (24 h) (Figure 1(a)). The
titratable acidity increased significantly ( p , 0.05)
with the increase in fermentation time in all batter
combinations from 0.17 to 1.05% (Figure 1(b)).
There is no significant decrease and increase in the pH
and titratable acidity between the batter combinations.
However, the sample C (4:1) showed higher acid

Table II. Microbial profile of finger millet–horse gram blended fermentation.

Batter combination

Total aerobic bacteria (log cfu/g

batter) Lactic bacteria (log cfu/g batter) Yeast (log cfu/g batter)

Fermentation time (h) A B C D A B C D A B C D

0 7.04 7.29 7.12 7.25 6.04 6.74 6.52 6.34 0.00 4.04 0.00 4.52
4 7.38 7.60 7.76 7.66 6.74 7.04 7.27 7.24 0.00 4.52 4.74 4.34
8 7.94 8.02 7.99 8.00 7.32 7.52 7.47 7.47 4.64 4.74 4.99 4.94
12 8.21 8.24 8.00 8.06 7.59 7.87 7.69 7.59 5.04 5.12 5.19 5.08
16 8.23 8.36 8.32 8.19 7.94 8.04 7.80 7.81 5.22 5.30 5.40 5.27
20 8.20 8.19 8.24 8.18 7.98 7.99 7.82 7.90 5.39 5.40 5.46 5.40
24 8.15 8.20 8.23 8.15 7.99 7.94 7.93 7.95 5.44 5.34 5.50 5.36

Note: Batter combinations (Finger millet: Horse gram) – A (2:1); B (3:1); C (4:1) and D (5:1).
 Nutritional value of finger millet based food (Indian dosa) by co-fermentation with horse gram flour 9

proteins from 0 to 12 h in all blended proportions

(Figure 4(a)). However, after 12 h, the soluble protein
content declined significantly ( p , 0.05). There is a
significant ( p , 0.05) difference in the soluble protein
content between the batter combinations (A, B, C and
D). A significant change in total free amino acids
was observed in all batter combinations (Figure 4(b)).
In B and C, amino acids increased till 16 h, whereas
in A and D, free amino acids increase was observed
till 24-h fermentation. A high amount of free amino
acids 1.1 g/100 g was observed in C (4:1) at 16-h
fermentation.
Results on the total free amino acids indicated that
the sample C (4:1 ratio blended batter) is found to be
optimum for dosa preparation. Hence, further analysis
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13

on changes in the amino acids composition during

fermentation was analysed at 8 h intervals with sample
C. Amino acid profile and essential amino acid
For personal use only.

Figure 1. (a) pH and (b) titratable acidity profile of co-fermenting

finger millet and horse gram. Batter combinations (Finger millet:
Horse gram) – A (2:1); B (3:1); C (4:1) and D (5:1). Each values
represent mean ^ SE of the mean (n ¼ 3).

content of 1.05 g lactic acid equivalents/100 g at the

end of fermentation as compared to other proportions.

Total soluble, reducing and non-reducing sugars

Co-fermentation of finger millet and horse gram
resulted in significant ( p , 0.05) changes in the profile
of available sugars between fermentation time and
batter combinations. The total soluble sugar, reducing
sugar and non-reducing sugars increased substantially
at 4 h followed by a significant ( p , 0.05) decrease
up to 20 h, which was further stabilized insignificantly
at 24 h in all the four batter combinations (Figure 2).

Starch content
Starch content of the batter decreased significantly
(P , 0.05) from 0 to 24 h fermentation in all
combinations (Figure 3). Sample D (5:1) showed
higher decrease in the starch content (25.52%)
followed by B (23.93%), C (23.26%) and A (21.30%)
at the end of fermentation.
Figure 2. Effect of co-fermentation on sugar content in finger
millet–horse gram blend. (a) Total soluble sugars; (b) reducing
Soluble proteins, total free amino acids and amino acid sugars and (c) non-reducing sugars. Batter combinations (Finger
profile during fermentation of batter millet: Horse gram) – A (2:1); B (3:1); C (4:1) and D (5:1). Results
are mean of three determinations ^ SD. Bars having different letters
Finger millet – horse gram batter fermentation led to a a– f (within fermentation time) and w–z (within batter combination)
significant ( p , 0.05) increase in soluble fraction of in particular sugar content are significantly different at p , 0.05.
 10 B. D. Palanisamy et al.

produced from the different blends showed that all the

samples were rated above average. However, the
sample C had the highest preference rating followed by
D, B and A. Appearance, texture and colour of the
product prepared from the four different proportions
scored more or less equal but their score values
differed in rating of taste, flavour and mouth feel.

Figure 3. Decrease in starch content during co-fermentation of
finger millet–horse gram blend. Batter combinations (Finger millet:
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
Discussion
The proximate composition of finger millet and horse
gram used in this study is comparable to the already
reported varieties. Crude fibre of finger millet is
distinctly higher than that of wheat and rice (Sripriya
et al. 1997), and the fat content of horse gram was

Horse gram) – A (2:1); B (3:1); C (4:1) and D (5:1). Values are

mean ^ SE of three independent determinations.
chemical scores calculated from relative amino acid
amount in sample/amino acid content in FAO (1991)
reference protein are presented in Table III. Fermen-
tation resulted in an increase in most of the individual
amino acids, while a decrease was also observed in
some amino acids. Total essential amino acids
increased from 48.25 to 48.73 g/100 g total amino
acids at the end of fermentation. Among the essential
amino acids, lysine increased significantly from 5.87
For personal use only.
found to be low.
From the above results, it can be seen that LAB and
yeast play a significant role throughout the fermenta-
tion process of finger millet – horse gram blend.
Microbiological studies of the batter showed the
predominance of LAB, which is on par with the
previous report of Chavan and Kadam (1989) on
cereal fermentation. The low yeast count indicates that
the fermentation was non-alcoholic and bubbles
formed during the fermentation of batter indicated
heterolactic fermentation. According to Steinkraus

to 6.73 g of amino acid/100 g of total amino acids. et al. (1967), initial bacterial counts in rice – black
Threonine, cysteine and isoleucine increased notably, gram mixture ranged from 103 to 105 cfu g2l, rising to
whereas other essential amino acid contents are more 108 – 109 cfu g21 after 20 –24 h of fermentation.
or less stabilized or decreased. In case of non-essential
amino acids, serine and arginine increased signifi-
cantly with a slight increase in alanine and other amino
acids decreased drastically. Essential amino acid score
of the batter was higher than the FAO amino acid
scoring pattern for all amino acids and scored higher
than 1.

Batter volume (%)

Fermentation of batter invariably leads to the
increased volume. As observed, there was no apparent
increase in batter volume in all proportions till 4 h of
fermentation. After 4 h, there was a little increase till
12 h. At the end of 16 h, batter volume increased
around 50% in 3:1, 4:1 and 5:1 proportions but there
was only 12.7% increase in 2:1 proportion. After that,
the batter volume started to decline in all blended
proportions, and a foul smell was noted on prolonged
fermentation above 24 h.

Sensory analysis
The scores of the sensory evaluation of dosa prepared
from different ratio blends of finger millet and horse
gram flour are given in Table IV. Sixteen hours
fermented batter was considered for dosa preparation
due to the sour aroma development and increased
batter volume. The consumer acceptability of the dosa
Figure 4. Changes in soluble proteins (a) and total free amino acids
(b) during co-fermentation of finger millet–horse gram flour. Batter
combinations (Finger millet: Horse gram) – A (2:1); B (3:1); C (4:1)
and D (5:1). Each values represent mean ^ SE of the mean (n ¼ 3).
Bars having different letters a–f (within fermentation time) and w–z
(within batter combination) in soluble protein content are
significantly different at p , 0.05.
 Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
For personal use only.

Table III. Amino acid profile and amino acid chemical score of finger millet –horse gram co-fermented batter (C (4:1 ratio blended)).

Amino acid profile* Essential amino acid chemical scores†

Fermented batter Fermented batter

Amino acids 0h 8h 16 h 24 h Essential amino acids FAO ref protein‡ 0h 8h 16 h 24 h

Essential amino acids

Threonine 4.27 4.29 4.33 4.31 Threonine 3.4 1.26 1.26 1.27 1.27
Valine 5.97 5.94 5.85 5.79 Valine 3.5 1.71 1.70 1.67 1.65
Cysteine 1.92 1.9 1.93 1.95 Methionine þ cysteine{ 2.5 1.88 1.84 1.86 1.87
Methionine 2.78 2.7 2.72 2.71 Isoleucine 2.8 1.50 1.52 1.56 1.63
Isoleucine 4.21 4.25 4.37 4.55 Leucine 6.6 1.49 1.48 1.47 1.45
Leucine 9.82 9.76 9.67 9.58 Phenylalanine þ tyrosine{ 6.3 1.53 1.51 1.51 1.51
Tyrosine 4.14 4.12 4.12 4.12 Histidine 1.9 1.35 1.35 1.35 1.35
Phenylalanine 5.47 5.42 5.39 5.4 Lysine 5.8 1.04 1.07 1.15 1.16
Histidine 2.57 2.56 2.56 2.56 Tryptophan 1.1 1.12 1.02 0.98 0.94
Lysine 5.87 6.23 6.66 6.73
Tryptophan 1.23 1.12 1.08 1.03
Total essential amino acids§ 48.25 48.29 48.68 48.73
Non-essential amino acids
Aspartic acid 7.15 6.89 5.92 4.56
Serine 5.64 5.87 5.96 6.21
Glutamic acid 22.59 20.45 19.78 17.34
Proline 5.78 5.45 4.23 4.56
Alanine 5.80 5.80 5.82 5.84
Arginine 4.82 4.89 5.12 5.45
Total amino acidsk 11.42 11.41 11.46 11.40

* Amino acid composition expressed in g of amino acid/100 g of total amino acids; † Amino acid chemical scores calculated from relative amino acid amounts in sample/amino acid content in FAO reference
protein; ‡ FAO reference protein pattern for preschool-age child (2 –5 years) (FAO 1991); { Pairs of amino acids considered physiologically substitute to one another (FAO 1991); § Total of essential amino
acids (g/100 g of total amino acids); k Total amino acids in g amino acids/100 g of dry matter calculated from protein content
Nutritional value of finger millet based food (Indian dosa) by co-fermentation with horse gram flour
11
 12 B. D. Palanisamy et al.
Table IV. Acceptability and sensory evaluation scores of dosa* (fermented finger millet–horse gram flour).
Characteristic of dosa‡ (pancake)
Samples† Appearance Texture Colour
A 5.3 ^ 0.95 5.1 ^ 0.88 5.9 ^ 0.74
B 5.1 ^ 1.10 4.8 ^ 0.92 5.3 ^ 0.95
C 5.1 ^ 0.88 5.7 ^ 0.95 5.0 ^ 0.82
D 4.8 ^ 0.79 5.4 ^ 1.07 4.8 ^ 1.03
* 7-point hedonic scale is as follows: 7-excellent, 6-very good, 5-good, 4-satisfactory, 3-fair, 2-poor, 1-very poor; † Batter combinations (Finger
millet: Horse gram) – A (2:1); B (3:1); C (4:1) D (5:1); ‡ Mean ^S.D. score of 10 determinations (10 panellists) of each three sets of batter;
{ Mean values bearing different letters a, b and c within the same column are significantly different ( p , 0.05) on the application of Duncan’s
multiple range test.
Acidification and leavening are the two important
changes that occur due to microbial activities in the idli
and dosa batters, which are involved in the production
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
of acid and gas from various carbohydrates. Inter-
actions between yeasts and lactobacilli are vital for
the metabolic activity of the batter. The acid and gas
required for leavening are produced exclusively by
heterofermentative LAB such as Leuconostoc mesenter-
oides, Streptoccoccus faecalis and Pediococcus cervisae
(Steinkraus et al. 1967). During dosa batter fermenta-
tion, L. mesenteroides, S. faecalis, Lactobacillus fermen-
tum and Bacillus amyloliquefaciens are the predominant
bacteria that are responsible for souring and leavening.
For personal use only.
In addition, yeasts such as Saccharomyces cerevisiae,
Debaryomyces hansenii and Trichosporon beigelli pro-
duced flavour along with the enzymes that help in the
saccharification of starch (Soni et al. 1986). In this
study, 10 different predominant lactic cultures were
found which are gram-positive cocci and rods and all
were catalase negative. Three strains were identified by
16S rRNA sequencing as Enterococcus durans (Gen-
Bank accession no. JF920299), Enterococcus faecium
(JF920300) and Lactobacillus plantarum (JF920301).
Identification of remaining strains is in progress.
Biochemical changes such as drop in pH and
increase in acidity, sugars, soluble proteins and amino
acids observed in this study are mainly attributed by
the activities of fermenting microbes. They secrete a
number of enzymes, which catalyze the hydrolysis of
carbohydrates, lipids, proteins and anti-nutritional
factors (Rolle 1998). A drop in pH and increase in
titratable acidity during co-fermentation of finger
millet and horse gram are in accordance with the
results of autofermentation of finger millet alone
(Sripriya et al. 1997). A pH range of 3.6 –4.1 is
evidently favourable for eliminating undesirable
microbial flora (coliforms) in fermented foods
(Chavan and Kadam 1989).
Decrease in sugars in all batter combinations could
be attributed to the utilization of sugars by fermenting
microorganisms that are generally highly active, and
this exploitation is higher than the rate of production.
After 20 h of fermentation, sugar content is stabilized
as the growth of microorganisms ceases due to acid
accumulation. Acids, especially lactic acids accumu-
late due to the conversion of sugars to acids. Similarly,
Taste Flavour Mouth feel Overall, acceptability{
4.6 ^ 1.17 4.8 ^ 0.92 4.9 ^ 0.99 5.08 ^ 1.00c
5.1 ^ 0.74 5.3 ^ 0.67 5.4 ^ 0.84 5.16 ^ 0.86c
6.5 ^ 0.71 5.7 ^ 0.82 6.1 ^ 0.88 5.70 ^ 0.97a
6.0 ^ 0.82 5.9 ^ 0.74 5.9 ^ 0.57 5.48 ^ 0.96b
Khetarpaul and Chauhan (1991) have reported a
significant reduction in the amount of total soluble,
reducing and non-reducing sugars and starch during
sequential fermentation of pearl millet (Pennisetum
typhoideum) using yeasts, lactobacilli. Reducing sugars
(as glucose) showed a steady decrease from 3.3 mg/g
in dry ingredients to 0.8 mg/g in 20 h in idli batter
fermentation (Desikachar et al. 1960). Hydrolysis
of starch by fermenting microbes that possess alpha
amylase reduces starch content in the fermented
product. In this study, starch content was decreased
nearly by 25% at the end of fermentation and
hydrolyzed to sugars and in turn converted to acids
and gas, which is evidenced by an increase in acidity
and microbial population. Due to the availability of the
excess of reducing sugars, the microbes grew faster
and multiplied in large numbers and brought about
an increase in acid levels. Sripriya et al. (1997)
reported that the decrease in starch content by 7.4%
during germination and fermentation of finger millets
may be due to fermenting carbohydrates (starch).
Increase in the soluble protein content may be
ascribed by an increase in the microbial enzyme
activity, protein hydrolysis and breakdown of tannins
and phytates (Chavan and Kadam 1989). Phytates
form a strong insoluble complex with proteins, and
their interaction is pH dependent and resists proteol-
ysis. Low pH in the fermentation medium solubilizes
the complex, and lactic and acetic acids enhance the
phytase activities, resulting in lowering of phytate and
thereby increasing the level of soluble proteins (Lopez
et al. 1983). Tannins form complex with enzymes of
digestive tract and also bind with proteins thereby
adversely affecting the utilization of proteins. Fermen-
tation reduces the tannin content, which might be due
to the action of microbial enzymes on tannin – protein
complex resulting in the release of proteins. Soluble
proteins and free amino acids increase is also due to
the hydrolysis of insoluble proteins by bacterial
proteases. The microorganisms are able to hydrolyze
proteins into usable amino acids and peptides. These
amino acids can be readily utilized by the microflora
for their metabolic activity. Additionally, during their
growth cycle, they can synthesize amino acids from
metabolic intermediates (Au and Fields 1981). When
the amino acid increment (promoted by the referred
 Nutritional value of finger millet based food (Indian dosa) by co-fermentation with horse gram flour
production and/or by protein hydrolysis) is superior to
the amino acid utilization, a final increase in free
amino acid levels is observed (Correia et al. 2010).
The total free amino acids were observed to increase
rapidly for about 4 – 5-fold during germination and
doubled at 18 h fermentation of finger millets (Sripriya
et al. 1997).
Significant changes in amino acids are due to the net
effect of LAB, yeast and fermenting conditions, which
determines the direction of changes in individual
amino acids. Methionine, the limiting amino acid in
legume and lysine, the limiting amino acid in cereals
are increased by co-fermentation. On mixing finger
millets and horse gram, the proteins complement one
another to produce protein of a better quality by
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
providing each other significant amounts of the
respective limiting amino acids. In this study, essential
amino acids such as cysteine, threonine and isoleucine
increased nearly one fold, whereas lysine increased 1.2-
fold. The degradation of prolamins into lower peptides
and free amino acids supplies amino groups, which
may be used through transamination to synthesize
lysine. Glutamic acid and proline have been implicated
in providing nitrogen for the synthesis of lysine and
other essential amino acids (Mbithi-Mwikya et al.
2000). These results are in accordance with cereal –
For personal use only.
legume (idli, dosa and dhokla), cereal –milk (tarhana
and kishk)-based fermented food products. The
proteins in the fermented batter contained 48.25 –
48.73% essential amino acids when compared to total
amino acids, and were higher than essential amino
acids (33.9%) in the FAO reference protein pattern
for children 2 –5 years old (FAO 1991).
Fermentation of idli batter appears to have a
significant effect on the increase of all essential
amino acids (Steinkraus et al. 1993). An increase of
10.6% (Steinkraus et al. 1967), 18.6% (Rao 1961) and
60.0% (Padhye and Salunkhe 1978) of methionine in
fermented idli batters over unfermented has been
reported. Dhokla is also similar to idli except for
the replacement of black gram with Bengal gram.
A significant improvement in the biological value and
net protein utilization of dhokla due to fermentation
have also been reported (Aliya and Geervani 1981).
The increase in the content of total free amino acids
and total free essential amino acids of fermented
tarhana (yoghurt– cereal mixture) was 57 and 93%,
respectively. The amino acids primarily responsible for
the increase were valine and tryptophan followed by
methionine, alanine, isoleucine þ leucine, phenyl-
alanine, arginine, proline and lysine (Erbas et al.
2005). Similarly, the protein content of kishk, a
fermented (milk– wheat mixture) Egyptian product is
also very high and in turn the amino acid content is
excellent. Kishk contains high concentrations of
phenylalanine, threonine, isoleucine, leucine, argi-
nine, valine, tyrosine and lysine, but it has low
amounts of tryptophan and sulphur-containing amino
acids (Morcos et al. 1973).
13
Another important factor in the fermentation of idli
or dosa batter is leavening and is essential for the
porosity of the steamed pudding. In the co-fermenta-
tion of finger millet and horse gram, microbes and
components present in the ingredients are responsible
for the evolution of gaseous products, which in turn
increases the batter volume. The batter of fermented
foods like idli or dosa is foamy in which gas molecules
are entrapped in a solid –liquid phase. These foams
are colloidal systems in which tiny air bubbles are
dispersed in continuous phase, wherein surface-active
agents can decrease the interfacial tension that
facilitates the gas –water interface system (Nisha et al.
2005). The soft spongy texture observed in the
leavened steamed idli or the puddings prepared from
the black gram are due to the presence of two
components: surface active protein (globulin) and
arabinogalactan (polysaccharide) in black gram
(Susheelamma and Rao 1979). Arabinogalactan is
known to stabilize the network of foam formed by the
surface-active protein and is essential for the rising of
the rice/black gram batters and for the porosity of the
steamed puddings. Further, arabinogalactan also holds
the leavening gases and prevents the disruption of
protein-foam or the spongy texture at cooking
temperatures. The compositional analysis of horse
gram non-starch polysaccharide indicates the presence
of arabinoxylan and xyloglucan as constituents of
hemicellulose which are similar to other legumes. This
leads to the rise in batter volume of finger millet – horse
gram blended batter similar to rice – black gram
blended batter.
In our present study, after 16-h fermentation, the
batter volume started to decline which might be due to
microbial interactions with ingredients. A fouling smell
also started to appear after 16-h fermentation due
to over fermentation, leading to the accumulation of
gaseous products and acids. Breakdown of starch
structure often leads to liquefaction of the batter,
leading to the loss of viscosity and gas retaining capacity
of the batter. Earlier, Iyer and Ananthanarayan (2008)
have reported that, on increasing the amylase concen-
tration above a certain level, a drastic fall in batter
viscosity occurred in idli batter, which is due to the fact
that the addition of amylase enzyme in excess amount is
known to cause liquefaction of starch. Nisha et al.
(2005) reported that after a certain period of fermenta-
tion the batter starts collapsing, and on long-term
storage, whey gets separated, resulting in idli with a hard
texture and over fermented flavour. Due to the air–
water interface, which is a high-energy interface, foam
collapses and these problems persist during storage.
During the preparation of dosa, a sol to gel
transformation occurs during the heating and within
a few minutes, a circular, semi-soft to crisp product
resembling a pancake, ready for consumption is
obtained. Dosa prepared from 4:1 ratio blended
proportion scores good acceptability; however, 2:1
and 3:1 proportion blend product scored very less due
 14 B. D. Palanisamy et al.
to the bean flavour of horse gram which was present in
high proportion and 5:1 ratio blended product rated
second due to high ragi floury taste and mouth feel.
Thus, the proportion of ingredients (finger millet and
horse gram) used for the preparation of fermented
food product is essential. The success of fermentation
depends on the acceptability of the products.
Fermentation leads to rapid acidification of the raw
material through the production of organic acids,
mainly lactic acid; in addition, certain aroma
compounds, bacteriocins, exopolysaccharides and
several enzymes are produced. As a result, they yield
non-toxic substrates (within limits) that often have
desirable spin-offs, such as enhanced taste, aroma,
shelf life, texture and nutritional value, which are
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
attractive and palatable to the human consumer
(Steinkraus 2002). According to Chavan and Kadam
(1989), during cereal fermentations several volatile
compounds were formed, which contribute to a
complex blend of flavours in the products. It has
been reported by Soni and Arora (2000) that yeasts
involved in fermentation not only contribute towards
gas production that results in good texture, but also
towards the sensory qualities of the idlis. Generally,
over fermentation gives way to more acid accumu-
lation resulting in sour taste and undesirable odour,
For personal use only.
which is not preferred for household purposes.
Conclusion
This work was conducted to study the biochemical and
microbial changes during co-fermentation of finger
millet and horse gram in different proportions for the
preparation of dosa, which plays a significant role in
maintaining the final product quality. On the basis of
the results, 16-h fermentation of 4:1 ratio blended
batter was found to be better compared to other
proportions. This fermented product (dosa) is
expected to contribute immensely as a protein source
and dietary supplement in developing countries like
India. However, further research works are warranted
to evaluate the availability of nutrients and other
nutritional requisites like vitamins and minerals by
in vitro and in vivo studies. Identification and
characterization of inhabiting microorganisms and
development of starter cultures for improved
product quality are being carried out. Efforts will be
made further to transfer this product to household
purposes.
Declarations of interest: Authors do not have any
conflict of interest.
References
Aliya S, Geervani P. 1981. An assessment of the protein quality and
vitamin B content of commonly used fermented products of
legumes and millets. J Sci Food Agric 32:837–842.
Antony U, Sripriya G, Chandra TS. 1996. Effect of fermentation on
the primary nutrients in finger millet (Eleusine coracana). J Agric
Food Chem 44:2616–2618.
AOAC. 2000. Official Methods of Analysis of the Association of
Official Agricultural Chemists. Washington, DC: Association of
Analytical Chemists. p 125–139.
Au PM, Fields ML. 1981. Nutritive quality of fermented sorghum.
J Food Sci 46:652–654.
Blandino A, Al-Aseeri ME, Pandiella SS, Cantero D, Webb C. 2003.
Cereal-based fermented foods and beverages. Food Res Int 36:
527 –543.
Bravo L, Siddhuraju P, Saura-Calixto F. 1999. Composition of
underexploited Indian pulses. Comparison with common
legumes. Food Chem 64:185–192.
Chavan JK, Kadam SS. 1989. Critical reviews in food science and
nutrition. Food Sci 28:348–400.
Correia S, Nunes A, Guedes S, Barros AS, Delgadillo I. 2010.
Screening of lactic acid bacteria potentially useful for sorghum
fermentation. J Cereal Sci 52:9–15.
Desikachar HSR, Radhakrishnamurthy R, Rao GR, Kadkol SB,
Srinivasan M, Subramanyan V. 1960. Studies on idli fermenta-
tion. Some accompanying changes in the batter. J Sci Ind Res 19:
168 –172.
Dubois MK, Gilles KA, Hamilton JK, Rebers PA, Smith F. 1956.
Colorimetric method for determination of sugars and related
substances. Anal Chem 28:350–356.
Duncan DB. 1955. Multiple range and multiple F-test. Biometrics
11:1 –42.
Erbas M, Ertugay MF, Erbas MO, Certel M. 2005. The effect of
fermentation and storage on free amino acids of tarhana. Int J
Food Sci Nutr 56:349–358.
Food and Agricultural Organization of The United Nations (FAO).
1991. Amino acid scoring pattern. In: Protein Quality
Evaluation. Vol. 51. Rome, Italy: FAO/WHO Food and
Nutrition paper. p 21– 25.
Gopalan C, Ramashastri BV, Balsubramanium SC. 2000. Nutritive
value of Indian foods. National Institute of Nutrition,
Hyderabad. p 47.
Hubert J, Berger M, Nepveu F, Paul F, Daydé J. 2008. Effects of
fermentation on the phytochemical composition and antioxidant
properties of soy germ. Food Chem 109:709–721.
Iyer BK, Ananthanarayan L. 2008. Effect of a-amylase addition on
fermentation of idli - a popular south Indian cereal—Legume-
based snack food. LWT Food Sci Technol 41:1053 –1059.
Khetarpaul N, Chauhan BM. 1991. Sequential fermentation
of pearl millet by yeasts and lactobacilli: changes in available
carbohydrate content. Food Chem 40:235–240.
Lopez Y, Gordon DT, Fields ML. 1983. Release of phosphorus from
phytate by natural lactic acid fermentation. J Food Sci 48:
935 –954.
Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. 1951. Protein
measurement with the Folin Phenol Reagent. J Biol Chem 193:
265 –276.
Magne C, Larher F. 1992. Higher sugar content of extracts interferes
with the calorimetric determination of amino acids and free
proline. Anal Biochem 200:115–118.
Malleshi NG, Hadimani NA. 1993. Nutritional and technological
characteristics of small millets and preparation of value added
products from them. In: Riley KW, Gupta SC, Seetharam A,
Mushonga JN, editors. Advances in Small Millets. New Delhi:
Oxford and IBH Publishing Co. p 270–287.
Mbithi-Mwikya S, Ooghe W, Van Camp J, Ngundi D, Huygebaert
A. 2000. Amino acid profiles after sprouting, autoclaving and
lactic acid fermentation of finger millet (Eleusine coracana) and
kidney beans (Phaseolus vulgaris L.). J Agric Food Chem 48:
3081–3085.
McCready RM, Guggolz J, Siliviera V, Owens HS. 1990.
Determination of starch and amylase in vegetables. Anal Chem
22:1156–1158.
 Nutritional value of finger millet based food (Indian dosa) by co-fermentation with horse gram flour
Morcos SR, Hegazi SM, El-Damhougy ST. 1973. Fermented foods
in common use in Egypt-I. The nutritive value of Kishk. J Sci
Food Agric 24:153–156.
Nisha P, Ananthanarayan L, Singhal RS. 2005. Effect of stabilizers
on stabilization of idli (traditional south Indian food) batter
during storage. Food Hydrocolloids 19(2):179– 186.
Nout MJR. 2009. Rich nutrition from the poorest – Cereal
fermentations in Africa and Asia. Food Microbiol 26(7):
685–692.
Onyango C, Noetzold H, Bley T, Henle T. 2004. Proximate
composition and digestibility of fermented and extruded uji from
maize-finger millet blend. LWT Food Sci Technol 37:827–832.
Padhye VW, Salunkhe DK. 1978. Biochemical studies on black gram
(Phaseolus mungo L.) III. Fermentation of black gram and rice
blend and its influence on the in vitro digestibility of the proteins.
J Food Biochem 2:327–347.
Rao MVR. 1961. Some observations on fermented foods. In:
Progress in Meeting Protein Needs of Infants and Preschool
Int J Food Sci Nutr Downloaded from informahealthcare.com by Penn State Berks Campus on 08/19/13
Children. Washington, DC: Publication No. 843. National
Academy of Sciences, National Research Council. p 291 –293.
Ravindran G. 1992. Seed protein of millets: Amino acid
composition, proteinase inhibitors and in vitro protein digest-
ibility. Food Chem 44:13– 17.
Ray PK. 1969. Toxic factor(s) in raw horse gram (Dolichos biflorus).
J Food Sci 6:207– 211.
Reddy NR, Pierson MD. 1994. Reduction in antinutritional and
toxic components in plant foods by fermentation. Food Res Int
27:281– 290.
For personal use only.
15
Rolle RS. 1998. Review: Enzyme applications for agro-processing
in developing countries: An inventory of current and potential
applications. World J Microbiol Biotechnol 14:611–619.
Somogyi M. 1952. Note on sugar determination. J Biol Chem 195:
19–25.
Soni SK, Arora JK. 2000. Indian fermented foods: Biotechnological
approaches. Food Processing: Biotechnological Application.
New Delhi, India: Asiatech Publishers Pvt. Ltd.
Soni SK, Sandhu DK, Vikhu KS, Karma N. 1986. Micro-
biological studies on dosa fermentation. Food Microbiol 3:
45–53.
Sripriya G, Antony U, Chandra TS. 1997. Changes in carbohydrate,
free amino acids, organic acids, phytate and HCl extractability of
minerals during germination and fermentation of finger millet
(Eleusine coracana). Food Chem 58(4):345–350.
Steinkraus KH, Ayres R, Olek A, Farr D. 1993. Biochemistry
of Saccharomyces. In: Steinkraus KH, editor. Handbook of
indigenous fermented foods. New York: Marcel Dekker.
p 517 –519.
Steinkraus KH, van Veen AG, Theibeau DB. 1967. Studies on idli -
an Indian fermented black gram rice food. Food Technol 21:
916 –919.
Steinkraus KH. 2002. Fermentations in world food processing.
Compr Rev Food Sci Food Safety 1:23– 32.
Susheelamma NS, Rao MVL. 1979. Effect of simple processing on
the properties of protein and polysaccharides from black gram.
J Food Technol 14:463–472.