Four different types of traditional starter culture called inocula and coded A, B, C and D as described below (Table 1) were used by six (6) experienced traditional cassava processors in Burkina Faso to ferment cassava dough for attiéké processing. The cassava dough got from the four different types of cassava dough and the attiéké obtained were sampled in duplicate by using a sterile stomacher bags placed in icebox with ices and transported in the laboratory of microbiology of Food Technology Department (DTA/IRSAT/CNRST) for analysis.

In total, six (6) samples of imported cassava dough from Côte d’Ivoire, six (6) samples of cassava dough fermented with roasted inoculums, four (4) samples of cassava dough fermented with blanched inoculums and two (2) samples of cassava dough fermented with soaked inoculums were collected for microbiological analyses

as lactic acid bacteria, molds and yeasts numeration, impurities, starch, pH and total acidity. And 18 sam- ples of attiéké obtained from these fermented cassava dough were also collected for proximate compound determination of moisture, total carbohydrates, proteins, lipids, ash, energy value, minerals contents and mycotoxins (Aflatoxin B₁, B₂, G₁, G₂ and ochratoxin A) analyses.

Ten gram (10 g) of each sample were dissolved in 50 ml of sterile peptoned buffered water and mixed. The pH was directly measured with a numeric pH-meter (WTW multi line P4). For Total acidity, 10 g of each sample are mixed with 50 ml of distilled water in an erlenmeyer. 10 ml of the dilution were then titrated against 0.1 N KOH using phenolphthalein as indicator; the total acidity was then calculated as a percentage of lactic acid.

The preparation of samples and tenfold dilutions for inoculation outs agar plates were carried out according to ISO-6887 [12] . For all numerations, 10 g of the samples were homogenized in a stomacher bags with 90 ml of sterile peptoned buffered water. The Tenfold serials dilution was prepared and spread-plated for microorganisms count. Yeasts and Molds were cultivated on Sabouraud-Chloramphenicol Agar (Oxoid LTD, Basingstore, Hamsphire, England) after incubated at 25˚C for 4 - 5 days and counted according to ISO 7954 standards [13] . Lactic acid bacteria were cultivated on De Man, Rogosa and Sharpe Agar (MRS, Merck 10660, Merck, Darmstadt, Germany) incubated anaerobically in an anaerobic conditions (anaerobic jar) at 37˚C, for 2 - 3 days and counted according to ISO 15214 standards [14] .

Proximate analysis of samples was conducted using the following conventional procedures described by the Association of Official Analytical Chemists [15] . Dry matter was determinate by drying at 105˚C ± 2˚C overnight; ash content by incineration at 550˚C for 12 h, crude protein (N × 6.25) by the Kjeldahl method after acid digestion; and crude fat content by Soxhlet extraction using n-hexane. Total carbohydrate content was determined by the phenol sulphuric acid method according to Tollier and Robin [16] and the values were expressed in g/100 g of attiéké. The starch content was determined using the colorimetric method described by Jarvis and Walker [17] . The energy value was calculated using the method described by Merrill and Watt [18] .

The contents of the minerals (Ca, Mg, Fe, Zn, Na, K, P) were determined after digesting of 0.5 g of sample using the Atomic Absorption Spectrophotometric method as outlined in the Association of Official Analytical Chemists Approved method [15] .

The analytical method used was based on liquid extraction with purification (5 g of sodium chloride + 125 ml of methanol extraction solution), followed by analysis by high-performance liquid chromatography with fluorescence detector (HPLC/FLD) and a photochemical post-column reactor according to ISO 16050 [19] .

For the determination of organic impurities 100 g of each sample were placed on a sieve (mesh=1.4 mm) and drained with water. The retained residue was dried at 105˚C ± 2˚C for about 10 min before weighing and expressed as the organic impurities.

For the determination of inorganic impurities, 100 g of each sample were shaken for 30 mn in a beaker containing water. The sediment obtained was dried at 105˚C ± 2˚C for 10 min before weighing and expressed as the inorganic impurities.

All the analyses were performed in triplicate. The data analysis was conducted using EXCELL and STATA version 2010. For physico-chemical and biochemical parameters, Simple Statistic Analysis was used to get means and standard deviations.

The minerals content of attiéké in relation with the dry matter are: phosphorus, from 150 to 524 mg/100g; potassium, from 120 to 445mg/100g; sodium from 217 to 885 mg/100g; calcium, from 42 to 116 mg/100g; magnesium, from 15 to 42 mg/100g; iron, from 7.05 to 14.29 mg/100g and zinc, from 0.18 to 0.73 mg/100g as shows in Table 2.

Any attiéké sample tested reveals a contamination with aflatoxin (B₁, B₂, G₁, and G2) or with ochratoxin A. Then, either this food has not been contaminated by the toxigenic microorganisms or attiéké analyzed have some properties which do not hallow the growth of those microorganisms.

The imported cassava fermented dough has generally the highest content in acidity. This could be explained by the species of microorganisms associate in cassava dough fermentation [9] . There is a significant difference of cassava fermented dough acidity among producers and also from one production to another. And, according to the nature of the inoculum used, there is also a significant difference of the cassava dough acidity and it microbial charge.

The major part of sample has similar values of acidity to the Côte d’Ivoire norm for attiéké [21] . Sotomey [20] had found for the Beninese’s attiéké, an acidity rate of 2.12 ± 0.21 and a pH of 5.03 ± 0.24. The difference of acidity among producers and productions observed could be explained by fermentation process mastery, transportation and storage conditions and also the nature of the inoculum [6] [22] [23] . Traoré [24] also proved that cassava variety influence it acidity value. The low acidity value of the bad cassava dough is in correlation with its weak load of fermented microorganisms. Acidity then mainly depends on fermentation process. The producers regulate the acidity of cassava dough to have acceptable value of attiéké acidity. The cassava dough acidity is then evaluated by tasting. This subjective evaluation is also changing from one day to another.

The impurities (organic and inorganic) constitute the third choice criteria of attiéké according to the consumers. Its frequency rate is about 36.9% [25] . Diancoumba and Sawadogo [26] identified peeling and washing as

^a: local sample from Ouagadougou; ^b: local sample from Orodara; ^c: local sample from Gaoua.

well as the burying underground of cassava dough as being the main process operation that govern the impurities rate. Indeed, during the peeling of the roots, the producers don't remove the central cylinder like Benin’s producers [20] which increase the organic impurities value. The weak value of inorganic impurities showed an important reduction comparatively to previous results [27] . Sotomey [20] found a similar value in average and the Côte d’Ivoire norm recommends some values ranged from 75% to 85%. Microorganisms counts of this study are less than those of the previous works [9] [11] [28] .

Starch is the main component of the total carbohydrates of attiéké. It is an energizing food which had to be balanced with other foods [29] . Sahoré and Nemlin [30] found content in proteins of 1.75% ± 0.01% and 1.70% ± 0.00% respectively from soft and bitter cassava variety, and the Côte d’Ivoire norm defines a value of 1% to 2% as the normal protein rate according to the dry matter. Sotomey [20] had found protein content of 1.87% ± 0.23% according to the dry matter. The weak value of protein content in attiéké is firstly due to the weakness content of cassava in protein [31] - [33] . Many factors can explain the protein content of attiéké: the variety of cassava used [30] , soil composition, cultural technics and the process as well. The increase of proteins content in attiéké therefore depends on yeasts (unicellular protein organism) in cassava dough [34] [35] . Djoulde et al. [34] demonstrated that Lactobacillus plantarum and Rhizopus oryzaeuse as starter contribute to increase about 10% ± 2% of protein content comparatively with the traditional fermentation method.

The content in lipids of attiéké is low and contributes a little to attiéké energizing value. Sotomey [20] has found lipids content of 0.8 ± 0.09 g/100g in attiéké in relation to the dry matter. And, Sahoré and Nemlin [30] found lipids content of 1.25% ± 0.03% from soft cassava variety attiéké and 1.40 ± 0.05 g/100g from bitter cassava attiéké. Attiéké color and gluing factor depends on the nature and the quantity of oil added during the transformation process. About 0.5 l of oil is added for 55 Kg of cassava dough. The highest content in lipids of attiéké analyzed is 22 times higher than the lowest one.

Thus, attiéké macronutrients values showed an important variation which does not facilitate a proper estimation of its daily contribution to human body needs. The fluctuation of the average values is as shown in the following figure (Figure 8).

The ashes values found are similar to those of Sotomey [20] who had found on average, ashes content of 0.68 ± 0.09 g/100g. The Côte d’Ivoire norm [21] requires ashes values less than 1.4 g/100g of dry matter. Total mineral salts content of attiéké is appreciable. There is an interrelationship between the content in ashes and the origin of the attiéké (cassava). So, on average, the attiéké of Gaoua has the strongest content (0.77 ± 0.01 g/100g) followed by those of Ouagadougou (0.47 ± 0.1 g/100g) and then, the imported one (0.32 ± 0.00 g/100g) and lastly, the attiéké of Orodara (0.2 ± 0.03 g/100g). Those results could be explained by cassava variety used for attiéké production or soil content in mineral in the regions where cassava in cultivated. Processing also influences ashes content in attiéké. In fact, Ouagadougou producers used the same variety of cassava for their production but they do not have the same value of ashes content. Favier [36] , Diallo et al. [7] already put out the impact of processing on the nourishing value of the final product. According to Favier [36] the peeling operation unit only could cause the loss of 37% of cassava ashes.

The analyzed attiéké samples contained neither ochratoxins nor aflatoxins. Yandju et al. [37] , found toxigenic moulds involved in the softening of cassava roots during low-moisture fermentation. But, Kastner et al. [41] did not found any contamination of aflatoxins among Côte d’Ivoire traditional inoculum. Only trace amounts of up to 0.2 mg/kg ochratoxins were present in some samples. The main toxigenic moulds are A. parasiticus, A. flavus,A. ochraceus, A. carbonarius,P. verrucosum, P. nordicum [39] - [41] .

The processes influence significantly attiéké major components. The washing process influences starch and then carbohydrates content of attiéké. Sahoré and Nemlin [30] had found a variation of 13% on the total carbohydrates content according to the process. In average, the content in proteins of the attiéké produced in Ouagadougou is relatively higher than both the imported one and the one from Gaoua and Orodara. And, besides Ouagadougou producers who use the same cassava variety there is an important variation of the protein content in cassava (Figure 6).

This fluctuation is also observed with the imported dough attiéké. The variation among producers (Figure 6) and from one production to another (Figure 7), points out the impact of the processing on protein content in attiéké. Sahoré and Nemlin [30] showed that from the raw material to the final product (attiéké), the reduction of the content in protein is about 3 to 5%.

As showed in Figure 8 it appears that any producer manages to maintain a steady value of lipids content in attiéké from one production to another. The content in lipids increases the energizing value of attiéké but also presents a risk of degradation of the nutritional value during the storage. There is an important variation of moister also among producers and from one production to another. The mainly processing that influence attiéké content in moister are pressing, drying and cooking. This fluctuation is due to the fact that the processing is still traditional [40] .

Many productions process influence attiéké component and therefore its nutritional value. According to the coefficient of variation, the parameters that are influenced the most by processes are impurities, lipids, ashes, protein, starch and carbohydrates as it is showed in the following picture (Figure 8).