Social Stratification in the Sikh Population of Punjab (India) Has a Genetic Basis: Evidence from Serological and Biochemical Markers

Abstract

Aim of the Study: The present study was planned to assess whether social stratification in the Sikh population inhabiting the northwest border Indian state of Punjab has any genetic basis. Subjects and Materials: Blood samples were collected randomly from a total of 2851 unrelated subjects belonging to 21 groups of two low-ranking Sikh scheduled caste populations, viz. Mazhabi and Ramdasi, and a high-ranking Jat Sikh caste population of Punjab. Methods: The genetic profile of Sikh groups was investigated using a total of nine serobiochemical genetic markers, comprising two blood groups (ABO, RH(D)) and a battery of seven red cell enzyme polymorphisms (ADA, AK1, ESD, PGM1, GLO1, ACP1, GPI), following standard serological and biochemical laboratory protocols. Genetic structure was studied using original allele frequency data and statistical measures of heterozygosity, genic differentiation, genetic distance, and genetic admixture. Results: Great heterogeneity was observed between Sikh scheduled caste and Jat Sikh populations, especially in the RH(D) blood group system, and distribution of ESD, ACP1, and PGM1 enzyme markers was also found to be significantly different between many of their groups. Genetic distance trees demonstrated little or no genetic affinities between Sikh scheduled caste and Jat Sikh populations; the Mazhabi and Ramdasi also showed little genetic relationship. Genetic admixture analysis suggested a higher element of autochthonous tribal extraction in the Ramdasi. Conclusions: The present study revealed much genetic heterogeneity in differently ranking Sikh caste populations of Punjab, mainly attributable to their different ethnic backgrounds, and provided a genetic basis to social stratification present in this religious community of Punjab, India.

Introduction

Located on the northwest border of India, Punjab has been the gateway of the country for entry of different racial/ethnic groups over millennia. Neolithic culture entered the Indian subcontinent from Baluchistan province in Pakistan and it was carried by Dravidian-speaking Mediterranean people who came into contact with late (Upper) Paleolithic hunting and gathering indigenous (pre-Dravidian) people. It spread from thence to Punjab, and excavations of Harappa, Mohenjodaro, and other cities provided evidence of a flourishing Indus valley or Harappan civilization in the third and second millennia B.C. But this Calcolithic civilization came to an abrupt end around 1500 B.C., as a result of foreign invasion when nomadic pastoral tribes, often called Aryan, crossed the Hindu Kush Mountains and entered this land. This marked the beginning of the next cultural phase, initially in Indus valley of Punjab, of the Indo-European-speaking Aryan invaders. In the great historic epic Rig-Veda, there is a story of invasion and overcoming of previous inhabitants, the Dasa (Mourant, 1983). When society was divided on basis of division of labor into Brahmin, Kshatriya, Vaish, and Shudra varnas, defeated people (the Dasa) were kept under the Shudra category, at the lowest rung of the social hierarchy, and were given the most hated occupations such as scavenging and cleaning of night soil. In the Indian constitution, the term scheduled caste has been used to describe such people.

The territorial extent of Punjab extends from 29°30′ to 32°32′ North latitude and from 73°55′ to 76°50′ East longitude, with an area of 50,362 km². The Ravi, Beas, and Sutlej rivers of Punjab define three natural geographical regions of the state (Fig. 1A), viz. the Majha (between Ravi and Beas), Doaba (between Beas and Sutlej), and Malwa (below Sutlej). Of these, the Malwa region is the biggest and most important. As per Census of India (2001), Punjab had a population of 24,289,296.

FIG. 1.

(A) District map of Punjab showing areas of study (shaded). (B) State map of India showing location of Punjab (shaded).

Historically, the people of Punjab are supposed to be progeny of a mixture of several proto- and post-Harappan invaders who entered this land from the West from time to time and of original autochthonous settlers of the area (Rose, 1919). The Sikh religion arose in Punjab in the 15th century and was founded by Guru Nanak Dev, first of the 10 gurus (masters) of Sikh. The Mazhabi and Ramdasi are the two most important endogamous scheduled caste populations of Sikh. Morphologically, members of both these castes have short stature, short and broad nose, dark skin color, and black eyes and hair—features that apparently suggest a higher element of indigenous (autochthonous) extraction in them (Jammu, 1996). They were ascribed to the Aryo-Dravidian physical type by Risley (1915). On the other hand, Jat Sikh are characterized by high stature and long and sharp nose, and the predominant color of skin is very light transparent brown and that of hair and eye is brown. They belong to the Indo-Aryan physical type (Risley, 1915).

Mazhabi

Like the Chuhra representing the lowest social stratum of Hindu society, the Mazhabi are at the bottom of the Sikh caste hierarchy. Traditionally, the male Mazhabi works as an agricultural laborer on contract basis in fields with Jat Sikh landlords and the female Mazhabi is engaged in doing menial jobs such as removing cow dung and stools, etc., from homes of such upper caste people in villages. Thus, their occupations provide opportunities for mingling of Mazhabi with Jat Sikh in daily life.

Ramdasi

Although a very low occupational caste, above only perhaps the Mazhabi, the Ramdasi are an integral part of Sikh society. Like Hindu Chamar, traditionally they are leather workers—taking a carcass away from habitation, removing its hide, eating some of the flesh, tanning the hide, dyeing it, or making shoes and thongs from it.

Jat Sikh

Jat Sikh, an agricultural caste, has its origin in the Jat, a nonautochthonous tribe that entered Punjab from its home on the river Oxus in Central Asia and first occupied the Indus valley. Much before the earliest Muslim invasions, the Jat had spread into Punjab proper, where they were firmly established in the beginning of the 11th century (Ibbetson, 1916). Those of the Jat who embraced the Sikh religion are known as Jat Sikh. Numerically, they are the single largest dominant endogamous caste population of Punjab.

Barring some observations on ABO and RH(D) blood groups, scheduled caste populations of Punjab are little represented on its genetic map. Therefore, the present study was planned to investigate the genetic constitution of two major lower social strata Sikh caste populations of Punjab, viz. The Mazhabi and Ramdasi, along with the Jat Sikh, a high ranking caste population (Table 1), using original data on two serological (ABO, RH(D)) and seven biochemical (adenosine deaminase [ADA], adenylate kinase locus 1 [AK1], esterase D [ESD], phosphoglucomutase locus 1 [PGM1], glyoxalase locus 1 [GLO1], acid phosphatase locus 1 [ACP1], and glucosephosphate isomerase [GPI]) polymorphisms. Use of only blood group and enzyme marker data is a limitation of the present study.

Table 1.

Sizes of the Studied Populations of Punjab (Census of India, 2001)

Region and districts	Mazhabi	Ramdasi	Jat Sikh
Majha
Gurdaspur	128,661	131,362	934,963
Amritsar	730,851	33,421	2,383,575
Doaba
Kapurthala	33,631	91,201	448,654
Nawanshahr	15,493	9,375	219,856
Hoshiarpur	47,538	38,429	574,862
Jallandhar	192,161	53,303	740,841
Malwa
Ferozepur	153,345	75,624	896,628
Muktsar	180,098	43,588	576,231
Bhatinda	188,948	91,680	876,426
Sangrur	146,565	313,554	1,400,149
Rupnagar	51,230	177,836	635,651

Subjects, Materials, and Methods

After obtaining informed consent, finger prick blood samples were collected from a total of 2851 randomly chosen, unrelated, healthy male and female Sikh subjects belonging to the Mazhabi (n=501), Ramdasi (n=517), and Jat Sikh (n=1833) caste populations inhabiting the Punjab (Table 2) during 1989-2003.

Table 2.

Sample Sizes of the Studied Populations of Punjab

Region and districts	Mazhabi (n=501)	Ramdasi (n=517)	Jat Sikh (n=1833)
Majha
Gurdaspur	—	—	176
Amritsar	—	—	181
Doaba
Kapurthala	—	—	180
Nawanshahr	—	—	175
Hoshiarpur	—	—	172
Jallandhar	—	—	177
Malwa
Ferozepur	114	86	140
Muktsar	123	82	142
Bhatinda	96	103	157
Sangrur	92	119	164
Rupnagar	76	127	169

Red cells were typed for ABO and RH(D) blood groups by the tube method, following standard serological techniques. Hemolysates were phenotyped for different red cell enzymes by the biochemical technique of horizontal gel electrophoresis as per protocols of Murch et al. (1986) for ADA and AK1, Wraxall and Stolorow (1986) for ESD and PGM1, Scott and Fowler (1982) for GLO1, Wraxall and Emes (1976) for ACP1, and Detter et al. (1968) for GPI.

Allele frequencies were calculated in the ABO system following Yasuda (1984), in RH(D) system by square root method, and in enzyme markers by gene counting method (Mourant et al., 1976). Deviations from Hardy-Weinberg equilibrium were studied by goodness-of-fit chi-square (χ²_HW) test using phenotype data, and for this, rare variants in PGM1 system and infrequent phenotypes in ACP1 system were grouped with the respective heterozygote. Degrees of freedom (df) was calculated by subtracting the number of independent variables required to construct expected data from the number of classes of phenotypes of the genetic system. Intergroup differences were studied by the contingency χ² test using the observed allele number data. Genetic structure was investigated using statistical measures of heterozygosity and genic differentiation as given by Nei (1973) and Weir and Cockerham (1984), respectively. To study genetic relationships, pairwise estimates of genetic distance (D) (Nei, 1972) were calculated, and dendrograms were constructed by the UPGMA and NJ methods (Sneath and Sokal, 1973) using Power Marker (Liu, 2005). Genetic admixture was estimated as described by Chakraborty (1986).

Results

The observed number of phenotypes along with goodness-of-fit chi-square (χ²_HW) test values for the nine studied markers in 5 groups each of Mazhabi and Ramdasi and 11 of Jat Sikh are shown in Tables 3 -5; allele frequencies are listed in Tables 6 -8. Barring a few groups in ABO, ADA, AK1, PGM1, and ACP1 systems, the overall distribution of markers in Sikh samples was in Hardy-Weinberg equilibrium (Tables 3 -5).

Table 3.

Distribution of Blood Groups and Red Cell Enzyme Polymorphisms in the Mazhabi Population of the Malwa Region of Punjab

		Mazhabi
Genetic marker	Phenotype	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
Blood groups
ABO
	O	38	43	30	20	14
	A1	25	23	17	23	20
	A2	2	6	1	7	—
	B	38	38	42	35	29
	A1,B	9	9	5	7	9
	A2,B	2	4	1	—	4
	Total	114	123	96	92	76
	χ²_HW (df=2)	1.65	1.63	1.27	5.14	0.69
RH(D)
	RH(D)+	110	117	91	87	74
	RH(D)−	4 (3.51)	6 (4.88)	5 (5.21)	5 (5.43)	2 (2.63)
	Total	114	123	96	92	76
Red cell enzymes
ADA
	1	84	93	81	71	62
	1,2	30	28	15	20	12
	2	—	1	—	1	2
	Total	114	122	96	92	76
	χ²_HW (df=1)	2.62	0.50	0.69	0.10	1.99
AK1
	1	73	95	78	75	60
	1,2	39	26	15	16	16
	2	2	2	3	1	—
	Total	114	123	96	92	76
	χ²_HW (df=1)	1.58	0.02	3.76	0.02	1.05
ESD
	1	56	70	51	41	39
	1,2	46	44	38	41	34
	2	12	9	7	10	3
	Total	114	123	96	92	76
	χ²_HW(df=1)	0.30	0.32	0.00	0.00	1.79
PGM1
	1	55	60	47	49	32
	1,2	46	48	45	34	36
	2	13	15	3	8	7
	1,3	—	—	1	—	1
	2,5	—	—	—	1	—
	Total	114	123	96	92	76
	χ²_HW (df=1)	0.50	1.20	4.36^a	0.24	0.64
GLO1
	1	7	8	9	12	7
	1,2	43	55	43	31	34
	2	64	60	44	49	35
	Total	114	123	96	92	76
	χ²_HW (df=1)	0.00	0.97	0.11	3.54	0.09
ACP1
	A	7	13	8	7	8
	A,B	53	44	52	42	36
	B	53	66	36	40	32
	A,C	—	—	—	1	—
	B,C	1	—	—	2	—
	Total	114	123	96	92	76
	χ²_HW (df=1)	1.97	1.81	3.25	1.39	0.21
GPI
	1	109	113	91	88	75
	1,3	5	10	5	4	1
	Total	114	123	96	92	76

Values in parentheses are percentages.

Statistically significant (p≤0.05)

Table 4.

Distribution of Blood Groups and Red Cell Enzyme Polymorphisms in the Ramdasi Population of the Malwa Region of Punjab

		Ramdasi
Genetic marker	Phenotype	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
Blood groups
ABO
	O	24	32	22	34	44
	A1	16	19	32	32	23
	A2	2	—	1	2	1
	B	32	25	42	43	47
	A1,B	9	5	5	8	12
	A2,B	3	1	1	—	—
	Total	86	82	103	119	127
	χ²_HW (df=2)	3.67	0.07	6.71^a	2.39	1.05
RH(D)
	RH(D)+	85	80	99	115	123
	RH(D)−	1 (1.16)	2 (2.44)	4 (3.88)	4 (3.36)	4 (3.15)
	Total	86	82	103	119	127
Red cell enzymes
ADA
	1	66	68	82	90	92
	1,2	17	12	20	27	34
	2	3	1	1	2	1
	Total	86	81	103	119	127
	χ²_HW (df=1)	1.85	0.31	0.03	0.00	1.28
AK1
	1	69	63	81	102	98
	1,2	17	18	22	16	24
	2	—	1	—	1	5
	Total	86	82	103	119	127
	χ²_HW (df=1)	1.03	0.05	1.47	0.18	4.35^a
ESD
	1	51	52	66	73	75
	1,2	29	24	35	42	46
	2	6	6	2	4	6
	Total	86	82	103	119	127
	χ²_HW (df=1)	0.44	1.74	1.18	0.48	0.10
PGM1
	1	40	32	57	50	44
	1,2	38	38	36	49	63
	2	8	11	9	20	19
	1,7	—	1	1	—	1
	Total	86	82	103	119	127
	χ²_HW (df=1)	0.06	0.03	0.70	1.73	0.30
GLO1
	1	6	7	8	10	15
	1,2	37	44	37	59	50
	2	43	31	58	50	62
	Total	86	82	103	119	127
	χ²_HW (df=1)	0.27	2.47	0.37	1.65	0.98
ACP1
	A	11	5	7	9	10
	A,B	38	28	44	41	65
	B	36	48	52	68	52
	A,C	1	—	—	—	—
	B,C	—	1	—	1	—
	Total	86	82	103	119	127
	χ²_HW (df=1)	0.01	0.05	0.32	0.49	2.83
GPI
	1	84	82	102	119	127
	1,3	2	—	1	—	—
	Total	86	82	103	119	127

Values in parentheses are percentages.

Statistically significant (p≤0.05).

Table 5.

Distribution of Blood Groups and Red Cell Enzyme Polymorphisms in the Jat Sikh Population of Punjab

		Jat Sikh
		Majha region		Doaba region				Malwa region
Genetic marker	Phenotype	Gurdaspur district	Amritsar district	Kapurthala district	Nawanshahr district	Hoshiarpur district	Jallandhar district	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
Blood groups
ABO
	O	53	67	73	72	59	73	46	51	42	42	61
	A1	35	28	32	32	40	26	32	25	39	16	34
	A2	6	3	8	9	5	3	5	5	1	1	6
	B	63	67	58	46	50	49	46	50	52	30	52
	A1,B	19	10	7	13	14	12	8	11	21	11	12
	A2,B	—	6	2	3	4	5	3	—	2	1	4
	Total	176	181	180	175	172	168	140	142	157	101	169
	χ²_HW (df=2)	3.86	17.53^a	0.61	2.20	2.78	19.35^a	1.12	2.36	5.10	5.90	1.67
RH(D)
	RH(D)+	156	159	163	154	158	143	120	120	136	87	153
	RH(D)−	20 (11.36)	22 (12.15)	17 (9.94)	21 (12.00)	14 (8.14)	25 (14.88)	20 (14.29)	22 (15.49)	21 (13.38)	14 (13.86)	16 (9.47)
	Total	176	181	180	175	172	168	140	142	157	101	169
Red cell enzymes
ADA
	1	122	140	129	136	135	149	111	103	129	115	125
	1,2	50	39	49	36	34	24	29	38	28	46	42
	2	4	2	2	3	3	4	—	—	—	3	1
	Total	176	181	180	175	172	177	140	141	157	164	168
	χ²_HW (df=1)	0.18	0.15	1.27	0.12	0.25	5.45^a	1.87	3.42	1.50	0.43	1.63
AK1
	1	142	149	141	137	137	128	111	104	133	124	134
	1,2	34	25	37	36	32	45	29	36	22	38	33
	2	—	7	2	2	3	4	—	1	2	2	1
	Total	176	181	180	175	172	177	140	141	157	164	168
	χ²_HW (df=1)	2.01	14.35^a	0.06	0.05	0.49	0.00	1.87	1.27	0.94	0.23	0.46
ESD
	1	108	117	120	97	116	112	91	90	105	100	107
	1,2	63	58	53	71	48	57	44	46	47	57	49
	2	5	6	7	7	8	8	5	5	5	7	12
	Total	176	181	180	175	172	177	140	141	157	164	168
	χ²_HW (df=1)	1.39	0.13	0.14	1.86	1.06	0.05	0.01	0.09	0.01	0.10	3.41
PGM1
	1	72	94	81	80	80	92	73	77	80	75	80
	1,2	86	70	66	78	69	69	58	54	65	74	75
	2	16	17	33	17	21	16	9	10	12	14	13
	1,7	2	—	—	—	2	—	—	—	—	—	—
	Total	176	181	180	175	172	177	140	141	157	163	168
	χ²_HW (df=1)	2.23	0.56	7.98^a	0.10	0.71	0.34	0.32	0.02	0.06	0.51	0.64
GLO1
	1	11	12	7	9	14	18	9	13	12	10	13
	1,2	76	72	79	70	75	77	53	55	71	71	58
	2	89	97	94	96	83	82	78	73	74	83	97
	Total	176	181	180	175	172	177	140	141	157	164	168
	χ²_HW (df=1)	0.98	0.08	3.80	0.69	0.27	0.00	0.00	0.32	0.80	1.04	1.06
ACP1
	A	17	21	22	25	25	31	17	19	20	20	16
	A,B	75	71	70	66	76	84	57	63	63	75	74
	B	83	88	81	84	69	62	63	56	73	65	77
	A,C	1	—	1	—	—	—	1	1	—	—	1
	B,C	—	1	6	—	2	—	2	2	1	4	—
	Total	176	181	180	175	172	177	140	141	157	164	168
	χ²_HW (df=1)	0.00	1.10	0.31	3.88^a	0.15	0.08	0.21	0.00	1.00	0.29	0.14
GPI
	1	176	179	178	175	171	170	140	141	157	164	167
	1,3	—	2	2	—	—	7	—	—	—	—	1
	1,7	—	—	—	—	1	—	—	—	—	—	—
	Total	176	181	180	175	172	177	140	141	157	164	168

Values in parentheses are percentages.

Statistically significant (p≤0.05).

Table 6.

Allele Frequencies in the Mazhabi Population of the Malwa Region of Punjab

	Mazhabi
Allele	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
ABOA1*	0.1624	0.1390	0.1242	0.1821	0.2195
ABOA2*	0.0150	0.0398	0.0094	0.0756	—
ABOB*	0.2435	0.2316	0.2971	0.2754	0.3285
ABOO*	0.5791	0.5896	0.5692	0.4669	0.4520
RHD*	0.8127	0.7791	0.7718	0.7669	0.8378
RHd*	0.1873	0.2209	0.2282	0.2331	0.1622
ADA1*	0.8684	0.8770	0.9219	0.8804	0.8947
ADA2*	0.1316	0.1230	0.0781	0.1196	0.1053
AK11*	0.8114	0.8780	0.8906	0.9022	0.8947
AK12*	0.1886	0.1220	0.1094	0.0978	0.1053
ESD1*	0.6930	0.7480	0.7292	0.6685	0.7368
ESD2*	0.3070	0.2520	0.2708	0.3315	0.2632
PGM11*	0.6842	0.6829	0.7292	0.7174	0.6645
PGM12*	0.3158	0.3171	0.2656	0.2772	0.3289
PGM13*	—	—	0.0052	—	0.0066
PGM15*	—	—	—	0.0054	—
GLO11*	0.2500	0.2886	0.3177	0.2989	0.3158
GLO12*	0.7500	0.7114	0.6823	0.7011	0.6842
ACP1A*	0.2939	0.2846	0.3542	0.3098	0.3421
ACP1B*	0.7017	0.7154	0.6458	0.6739	0.6579
ACP1C*	0.0044	—	—	0.0163	—
GPI1*	0.9781	0.9593	0.9740	0.9783	0.9934
GPI3*	0.0219	0.0407	0.0260	0.0217	0.0066

Table 7.

Allele Frequencies in the Ramdasi Population of the Malwa Region of Punjab

	Ramdasi
Allele	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
ABOA1*	0.1564	0.1598	0.2129	0.1893	0.1471
ABOA2*	0.0215	—	0.0110	0.0158	0.0066
ABOB*	0.2946	0.2109	0.2862	0.2507	0.2655
ABOO*	0.5274	0.6293	0.4899	0.5442	0.5808
RHD*	0.8922	0.8438	0.8029	0.8167	0.8225
RHd*	0.1078	0.1562	0.1971	0.1833	0.1775
ADA1*	0.8663	0.9136	0.8932	0.8697	0.8583
ADA2*	0.1337	0.0864	0.1068	0.1303	0.1417
AK11*	0.9012	0.8780	0.8932	0.9244	0.8661
AK12*	0.0988	0.1220	0.1068	0.0756	0.1339
ESD1*	0.7616	0.7805	0.8107	0.7899	0.7717
ESD2*	0.2384	0.2195	0.1893	0.2101	0.2283
PGM11*	0.6860	0.6280	0.7330	0.6261	0.5984
PGM12*	0.3140	0.3659	0.2621	0.3739	0.3976
PGM17*	—	0.0061	0.0049	—	0.0040
GLO11*	0.2849	0.3537	0.2573	0.3319	0.3150
GLO12*	0.7151	0.6463	0.7427	0.6681	0.6850
ACP1A*	0.3547	0.2317	0.2816	0.2479	0.3346
ACP1B*	0.6395	0.7622	0.7184	0.7479	0.6654
ACP1C*	0.0058	0.0061	—	0.0042	—
GPI1*	0.9837	1.0000	0.9951	1.0000	1.0000
GPI3*	0.0163	—	0.0049	—	—

Table 8.

Allele Frequencies in the Jat Sikh Population of Punjab

	Jat Sikh
	Majha region		Doaba region				Malwa region
Allele	Gurdaspur district	Amritsar district	Kapurthala district	Nawanshahr district	Hoshiarpur district	Jallandhar district	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
ABOA1*	0.1649	0.1115	0.1157	0.1362	0.1710	0.1188	0.1567	0.1349	0.2109	0.1399	0.1464
ABOA2*	0.0296	0.0136	0.0342	0.0383	0.0243	0.0133	0.0307	0.0283	0.0061	0.0074	0.0288
ABOB*	0.2676	0.2610	0.2095	0.1933	0.2198	0.2150	0.2318	0.2446	0.2718	0.2278	0.2247
ABOO*	0.5379	0.6139	0.6406	0.6322	0.5849	0.6529	0.5808	0.5922	0.5112	0.6249	0.6001
RHD*	0.6629	0.6514	0.6927	0.6536	0.7147	0.6142	0.6220	0.6064	0.6343	0.6277	0.6923
RHd*	0.3371	0.3486	0.3073	0.3464	0.2853	0.3858	0.3780	0.3936	0.3657	0.3723	0.3077
ADA1*	0.8352	0.8812	0.8528	0.8800	0.8837	0.9096	0.8964	0.8652	0.9108	0.8415	0.8690
ADA2*	0.1648	0.1188	0.1472	0.1200	0.1163	0.0904	0.1036	0.1348	0.0892	0.1585	0.1310
AK11*	0.9034	0.8923	0.8861	0.8857	0.8895	0.8503	0.8964	0.8652	0.9172	0.8720	0.8958
AK12*	0.0966	0.1077	0.1139	0.1143	0.1105	0.1497	0.1036	0.1348	0.0828	0.1280	0.1042
ESD1*	0.7926	0.8066	0.8139	0.7571	0.8140	0.7938	0.8071	0.8014	0.8185	0.7835	0.7827
ESD2*	0.2074	0.1934	0.1861	0.2429	0.1860	0.2062	0.1929	0.1986	0.1815	0.2165	0.2173
PGM11*	0.6591	0.7127	0.6333	0.6800	0.6715	0.7147	0.7286	0.7376	0.7166	0.6871	0.6994
PGM12*	0.3352	0.2873	0.3667	0.3200	0.3227	0.2853	0.2714	0.2624	0.2834	0.3129	0.3006
PGM17*	0.0057	—	—	—	0.0058	—	—	—	—	—	—
GLO11*	0.2784	0.2652	0.2583	0.2514	0.2994	0.3192	0.2536	0.2872	0.3025	0.2774	0.2500
GLO12*	0.7216	0.7348	0.7417	0.7486	0.7006	0.6808	0.7464	0.7128	0.6975	0.7226	0.7500
ACP1A*	0.3125	0.3122	0.3194	0.3314	0.3663	0.4124	0.3286	0.3617	0.3280	0.3506	0.3185
ACP1B*	0.6847	0.6851	0.6611	0.6686	0.6279	0.5876	0.6607	0.6277	0.6688	0.6372	0.6786
ACP1C*	0.0028	0.0027	0.0195	—	0.0058	—	0.0107	0.0106	0.0032	0.0122	0.0030
GPI1*	1.0000	0.9945	0.9944	1.0000	0.9971	0.9802	1.0000	1.0000	1.0000	1.0000	0.9970
GPI3*	—	0.0055	0.0056	—	—	0.0198	—	—	—	—	0.0030
GPI7*	—	—	—	—	0.0029	—	—	—	—	—	—

Genetic variation

The frequency of ABO*A1 varied from 0.1242 to 0.2195 in the Mazhabi (Table 6), from 0.1471 to 0.2129 in the Ramdasi (Table 7), and from 0.1115 to 0.2109 in the Jat Sikh (Table 8). The range of ABO*B frequency observed was 0.2316-0.3285, 0.2109-0.2946, and 0.1933-0.2718, respectively. ABO*A2 was found absent from one group each of the Mazhabi and Ramdasi, but it was present in every group of Jat Sikh. The χ² comparisons between scheduled caste and Jat Sikh groups (Tables 9 and 10) revealed significant heterogeneity in about one-third of cases.

Table 9.

Statistically Significant Intergroup χ ² Test Values Between the Mazhabi and Jat Sikh Populations of Punjab

		Jat Sikh
		Majha region		Doaba region				Malwa region
Locus	Mazhabi	Gurdaspur district	Amritsar district	Kapurthala district	Nawanshahr district	Hoshiarpur district	Jallandhar district	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
ABO (df=2)
	Bhatinda District				7.91^*	6.22^*				6.22^*
	Sangrur District	7.05^*	17.41^****	16.19^****	13.50^***	6.77^*	19.46^****	6.08^*	8.62^*		11.11^**	9.06^*
	Rupnagar District		12.93^***	15.72^****	15.36^***	8.65^*	17.55^****	7.03^*	7.84^*		10.49^**	9.64^*
ESD (df=1)
	Ferozepur District	7.39^***	9.98^***	11.41^****		11.18^****	7.60^***	8.87^***	7.96^***	11.59^****	5.82^**	5.78^**
	Muktsar District									4.09^*
	Bhatinda District		4.38^*	5.31^*		5.22^*		3.97^*		5.62^**
	Sangrur District	9.93^***	12.76^****	14.30^****	4.77^*	14.04^****	10.16^***	11.45^****	10.46^***	14.43^****	8.15^***	8.11^***
	Rupnagar District									4.14^*
PGM1 (df=1)
	Bhatinda District			5.17^*
	Sangrur District			4.36^*
ADA (df=1)
	Bhatinda District	8.02^***	5.53^**								6.97^***
AK1 (df=1)
	Ferozepur District	10.17^***	7.64^***	6.36^**	6.19^**	6.88^***		7.46^***		13.30^****		8.12^***
ACP1 (df=1)
	Ferozepur District						7.77^***
	Muktsar District					4.92^*	10.30^***		4.22^*		3.89^*
	Sangrur District						4.64^*

0.05>p>0.02.

0.02>p>0.01.

***

0.01>p>0.001.

****

p<0.001.

Table 10.

Statistically Significant Intergroup χ ² Test Values Between the Ramdasi and Jat Sikh Populations of Punjab

		Jat Sikh
		Majha region		Doaba region				Malwa region
Locus	Ramdasi	Gurdaspur district	Amritsar district	Kapurthala district	Nawanshahr district	Hoshiarpur district	Jallandhar district	Ferozepur district	Muktsar district	Bhatinda district	Sangrur district	Rupnagar district
ABO (df=2)
	Ferozepur District			6.60^*	7.35^*		7.61^*
	Muktsar District			7.33^*	6.89^**					6.77^*
	Bhatinda District		11.83^***	12.39^***	11.09^**	6.96^*	14.86^***	7.32^*			7.91^**	6.31^*
	Sangrur District		7.40^*				8.05^*
PGM1 (df=1)
	Muktsar District						3.91^*	4.90^*	5.89^**	3.92^*
	Bhatinda District			5.88^**
	Sangrur District		4.94^*				5.13^*	6.23^**	7.46^***	5.08^*
	Rupnagar District		8.75^***		4.28^*		8.98^***	10.15^***	11.73^****	8.78^***	4.92^*	6.53^**
ADA (df=1)
	Muktsar District	5.65^**									4.83^*
	Rupnagar District						3.92^*			3.87^*
AK1 (df=1)
	Sangrur District						7.40^***		4.69^*		3.99^*
	Rupnagar District									3.87^*
GLO1 (df=1)
	Muktsar District		4.26^*	4.99^*	5.73^**			5.02^*				5.82^**
	Sangrur District				4.51^*							4.59^*
ACP1 (df=1)
	Muktsar District			5.13^*	4.64^*	9.07^***	14.88^****	4.72^*	8.16^***	4.48^*	7.83^***
	Bhatinda District					4.71^*	9.63^***		4.08^*
	Sangrur District			4.79^*	4.25^*	9.26^***	16.12^****	4.30^*	8.14^***	4.05^*	7.81^***

0.05>p>0.02.

0.02>p>0.01.

***

0.01>p>0.001.

****

p<0.001.

The incidence of RH(D)−phenotype was found to be much lower in both Mazhabi (range: 2.63%-5.43%; Table 3) and Ramdasi (range: 1.16%-3.88%; Table 4) compared with Jat Sikh (range: 8.14%-15.49%; Table 5). Evidently, both Sikh scheduled caste populations of Punjab are characterized by comparatively low occurrence of this phenotype. In terms of allele frequency also, great differences were observed as RH*d was found in rather low ranges in both the Mazhabi (0.1622-0.2331; Table 6) and the Ramdasi (0.1078-0.1971; Table 7) compared with the Jat Sikh (0.2853-0.3936; Table 8). The χ² test between 11 groups of Jat Sikh and 10 of two scheduled castes was found to be statistically significant in almost all cases, save a few comparisons between the Jat Sikh of the Hoshiarpur district and the Mazhabi of the Muktsar, Bhatinda, and Sangrur districts and that between the Jat Sikh of Kapurthala district and the Mazhabi of Sangrur district. This overwhelmingly demonstrated conclusive heterogeneity in RH(D) distribution in upper- and lower-ranking Sikh caste populations of the Punjab.

As for polymorphic red cell enzyme systems, the range of ADA*2 frequency noted in Jat Sikh groups (0.0892-0.1648; Table 8) was found to be higher than that observed in groups of both Mazhabi (0.0781-0.1316; Table 6) and Ramdasi (0.0864-0.1417; Table 7). Apparently, Jat Sikh showed comparatively higher incidence of the allele than both scheduled castes, but statistically few χ² test values showed significant differences (Tables 9 and 10). Mazhabi showed somewhat higher incidence of AK1*2 (range: 0.0978-0.1886; Table 6) compared with both Ramdasi (range: 0.0756-0.1339; Table 7) and Jat Sikh (range: 0.0828-0.1497; Table 8), albeit statistically few χ² comparisons between groups of scheduled castes and Jat Sikh revealed appreciable differences in AK1 (Tables 9 and 10). Mazhabi showed ESD*2 frequency in a range (0.2520-0.3315; Table 6) that was much higher compared with that observed in Jat Sikh (0.1815-0.2429; Table 8), and intergroup comparisons revealed appreciable differences in half the cases (Table 9). The range observed in Ramdasi (0.1893-0.2384; Table 7) was almost identical with that of Jat Sikh and no heterogeneity was observed between them (Table 10).

In PGM1 besides three common phenotypes, three rare variants were found (Tables 3 -5). Ramdasi depicted somewhat a higher frequency range of PGM1*2 (0.2621-0.3976; Table 7) compared with both Mazhabi (0.2656-0.3289; Table 6) and Jat Sikh (0.2624-0.3667; Table 8). Rare alleles PGM1*3 and PGM1*5 were confined to Mazhabi (Table 6), whereas PGM1*7 was present in both Ramdasi and Jat Sikh (Tables 7 and 8). Almost all statistically significant χ² differences observed were between groups of Ramdasi and Jat Sikh (Tables 9 and 10). Similarly, in GLO1 it was the Ramdasi groups, which showed higher frequency range of GLO1*1 (0.2573-0.3537, Table 7) compared with the Mazhabi (0.2500-0.3177; Table 6) and the Jat Sikh (0.2500-0.3192; Table 8) groups. Few Ramdasi groups showed statistically significant differences with Jat Sikh groups for this marker (Table 10).

The Jat Sikh showed a consistently high frequency of ACP1*A (range: 0.3122-0.4124; Table 8) compared with both the Mazhabi (range: 0.2846-0.3542; Table 6) and the Ramdasi (range: 0.2317-0.3547; Table 7). Of 10 scheduled caste groups, only two Mazhabi and three Ramdasi showed presence of infrequent ACP1*C, found present in 9 of 11 Jat Sikh groups. Some significant differences were observed, especially between Ramdasi and Jat Sikh groups (Tables 9 and 10). Half of the Sikh groups studied showed presence of GPI variant 1,3, detected in all five Mazhabi, two Ramdasi, and four Jat Sikh groups (Tables 3 -5), revealing that the variant was frequent in the Sikh scheduled castes rather than the Jat Sikh.

Thus, in the present Sikh samples from Punjab, besides RH(D) blood group system, which showed almost complete heterogeneity between Sikh scheduled caste and Jat Sikh populations, distribution of ESD, ACP1, and PGM1 enzyme markers was also found to be significantly different between many of their groups.

Heterozygosity

Table 11 shows that there was great interlocus variation in heterozygosity (h) estimates in different Sikh populations. This heterogeneity in heterozygosity values could simply be a genetic consequence of population structure or it may indicate an effect of natural selection operating in the biosphere of Punjab. Because h is a measure of genic variation, in Sikh of Punjab, ABO was found to be the most variable marker, followed by ACP1, PGM1, GLO1, ESD, RH(D), ADA, AK1, and GPI (Table 11). On the other hand, average heterozygosity (H) estimates by population group (Table 12) were rather homogeneous in the Mazhabi (range: 0.3331-0.3528), the Ramdasi (range: 0.3118-0.3390), and the Jat Sikh (range: 0.3282-0.3485), suggesting a similar extent of genic variation in them.

Table 11.

Heterozygosity (h) in Different Sikh Populations of Punjab: Estimates by Locus

	Heterozygosity (h)
Locus	Mazhabi (5 groups)	Ramdasi (5 groups)	Jat Sikh (11 groups)
ABO	0.6071	0.5904	0.5636
RH(D)	0.3260	0.2728	0.4515
ADA	0.1975	0.2100	0.2175
AK1	0.2160	0.1910	0.2003
ESD	0.4057	0.3394	0.3225
PGM1	0.4243	0.4497	0.4231
GLO1	0.4141	0.4244	0.3992
ACP1	0.4372	0.4119	0.4556
GPI	0.0454	0.0084	0.0066
Average heterozygosity (H)	0.3415	0.3220	0.3378

Table 12.

Average Heterozygosity (H) in Different Sikh Populations of Punjab: Estimates by Population

Population	District	Average heterozygosity (H)
Mazhabi
	Ferozepur	0.3461
	Muktsar	0.3398
	Bhatinda	0.3331
	Sangrur	0.3528
	Rupnagar	0.3357
	Mean	0.3415
Ramdasi
	Ferozepur	0.3234
	Muktsar	0.3118
	Bhatinda	0.3142
	Sangrur	0.3215
	Rupnagar	0.3390
	Mean	0.3220
Jat Sikh
	Gurdaspur	0.3472
	Amritsar	0.3282
	Kapurthala	0.3369
	Nawanshahr	0.3369
	Hoshiarpur	0.3371
	Jallandhar	0.3444
	Ferozepur	0.3296
	Muktsar	0.3449
	Bhatinda	0.3299
	Sangrur	0.3485
	Rupnagar	0.3319
	Mean	0.3378

Genic differentiation

Estimates of F_ST in Sikh samples are presented in Table 13, which show that there was great heterogeneity in its values over loci. The average value of F_ST over eight diallelic loci was found to be 0.004539 in Mazhabi, 0.006618 in Ramdasi, and 0.004408 in Jat Sikh. This demonstrated that the Punjab groups of Jat Sikh as well as Mazhabi were comparatively less differentiated genetically than Ramdasi groups.

Table 13.

Estimates of F _ST in Different Sikh Populations of Punjab

	F _ST
Locus	Mazhabi (5 groups)	Ramdasi (5 groups)	Jat Sikh (11 groups)
RH(D)	0.004539	0.007089	0.005067
ADA	0.003544	0.003922	0.005327
AK1	0.009945	0.004174	0.003104
ESD	0.004331	0.001655	0.001833
PGM1	0.002704	0.010443	0.004313
GLO1	0.002914	0.005448	0.002480
ACP1	0.003132	0.010745	0.003654
GPI	0.005198	0.009465	0.009487
Average	0.004539	0.006618	0.004408

Nei's genetic distance

In the dendrogram constructed using the UPGMA method (Fig. 2), there were two distinct subclusters—one comprising solely of 11 Jat Sikh groups and other of all 10 groups of Mazhabi and Ramdasi. This bimodal clustering demonstrated that there were no genetic affinities between Jat Sikh and scheduled caste populations of Punjab, suggesting their diverse genetic compositions. In the latter subcluster, first three Ramdasi groups separate at an early stage of evolution and form a sub-sub cluster and similarly four Mazhabi groups are placed together in a sub-sub-sub cluster at the bottom of the tree. Evidently, genetic distinction between two Sikh scheduled caste populations was also discernible to a large extent from this dendrogram.

FIG. 2.

Dendrogram of 21 groups of Mazhabi, Ramdasi, and Jat Sikh populations of Punjab constructed from Nei's D matrix using the UPGMA method of clustering. JS-J, Jat Sikh, Jallandhar District; JS-B, Jat Sikh, Bhatinda District; JS-A, Jat Sikh, Amritsar District; JS-F, Jat Sikh, Ferozepur District; JS-M, Jat Sikh, Muktsar District; JS-S, Jat Sikh, Sangrur District; JS-G, Jat Sikh, Gurdaspur District; JS-H, Jat Sikh, Hoshiarpur District; JS-K, Jat Sikh, Kapurthala District; JS-N, Jat Sikh, Nawanshahr District; JS-R, Jat Sikh, Rupnagar District; RA-R, Ramdasi, Rupnagar District; RA-M, Ramdasi, Muktsar District; RA-S, Ramdasi, Sangrur District; RA-B, Ramdasi, Bhatinda District; RA-F, Ramdasi, Ferozepur District; MA-R, Mazhabi, Rupnagar District; MA-F, Mazhabi, Ferozepur District; MA-M, Mazhabi, Muktsar District; MA-B, Mazhabi, Bhatinda District; MA-S, Mazhabi, Sangrur District.

There were three subclusters in a dendrogram constructed using the NJ method (Fig. 3). First, Jat Sikh of Gurdaspur district separated in a single-line subcluster and another six groups of this caste form the second subcluster. In the third subcluster, two Jat Sikh groups delineated at an early stage of evolution, followed by the remaining two groups of this caste from Rupnagar and Hoshiarpur districts. The remainder of this subcluster encompassed all 10 groups of Mazhabi and Ramdasi. Thus, the NJ tree revealed little genetic affinity either between Jat Sikh and Sikh scheduled caste population groups or between groups of Mazhabi and Ramdasi, who tend to cluster separately.

FIG. 3.

Dendrogram of 21 groups of Mazhabi, Ramdasi, and Jat Sikh populations of Punjab constructed from Nei's D matrix using the NJ method of clustering. JS-G, Jat Sikh, Gurdaspur District; JS-A, Jat Sikh, Amritsar District; JS-B, Jat Sikh, Bhatinda District; JS-F, Jat Sikh, Ferozepur District; JS-S, Jat Sikh, Sangrur District; JS-M, Jat Sikh, Muktsar District; JS-J, Jat Sikh, Jallandhar District; JS-N, Jat Sikh, Nawanshahr District; JS-K, Jat Sikh, Kapurthala District; JS-R, Jat Sikh, Rupnagar District; JS-H, Jat Sikh, Hoshiarpur District; MA-B, Mazhabi, Bhatinda District; MA-M, Mazhabi, Muktsar District; RA-B, Ramdasi, Bhatinda District; MA-S, Mazhabi, Sangrur District; MA-F, Mazhabi, Ferozepur District; MA-R, Mazhabi, Rupnagar District; RA-F, Ramdasi, Ferozepur District; RA-R, Ramdasi, Rupnagar District; RA-S, Ramdasi, Sangrur District; RA-M, Ramdasi, Muktsar District.

Genetic admixture

The contention “the Indian lower castes are genetically more associated with the tribal populations than to the higher castes” (Thanseem et al., 2006) was evaluated in the context of Sikh lower castes of Punjab belonging to the Aryo-Dravidian physical type (Risley, 1915). For this analysis, a conglomerate of 16 autochthonous tribes of Central India, geographically nearest to Punjab, inhabiting states of Orissa, Madhya Pradesh, and Maharashtra (Fig. 1B) (Das et al., 1996), representing the Dravidian physical type and the present Jat Sikh of Punjab representing the Indo-Aryan physical type, “which approaches most closely to that ascribed to the traditional Aryan colonists of India” (Risley, 1915), were considered as parental populations of Mazhabi and Ramdasi.

Using average allele frequencies of the four common markers, viz. ABO, RH(D), ESD, and ACP1, in the Mazhabi, Ramdasi, and Jat Sikh populations of Punjab and Central Indian tribes (Table 14), genetic admixture estimates fitting a dihybrid model were calculated and the results are presented in Table 15. In case of Mazhabi, the major contribution came from Jat Sikh (67.42%), whereas in Ramdasi it was from Central Indian tribes (75.61%). Like Ramdasi, tribal contribution to Mazhabi gene pool was expected to be high, but apparently it has been offset by possible gene flow from Jat Sikh to the latter scheduled caste population because of their close proximity. Genetic admixture analysis, therefore, provided some evidence that lower strata of the Sikh population of the Punjab are more associated with Indian tribes than the Jat Sikh.

Table 14.

Average Allele Frequencies in the Mazhabi, Ramdasi, and Jat Sikh Populations of Punjab and Central Indian Tribes of Orissa, Madhya Pradesh, and Maharashtra

Marker	Allele	Mazhabi (5 groups)	Ramdasi (5 groups)	Jat Sikh (11 groups)	Central Indian tribes (16 groups) ^a
ABO		(501)	(517)	(1761)	(1651)
	ABOA1*	0.1654	0.1731	0.1461	0.1753
	ABOA2*	0.0280	0.0110	0.0231	0.0352
	ABOB*	0.2752	0.2616	0.2334	0.2298
	ABOO*	0.5314	0.5543	0.5974	0.5597
RH(D)		(501)	(517)	(1761)	(1651)
	RHD*	0.7937	0.8356	0.6520	0.9025
	RHd*	0.2063	0.1644	0.3480	0.0975
ESD		(501)	(517)	(1831)	(1612)
	ESD1*	0.7151	0.7829	0.7974	0.6710
	ESD2*	0.2849	0.2171	0.2026	0.3290
ACP1		(501)	(517)	(1831)	(1612)
	ACP1A*	0.3169	0.2901	0.3401	0.2210
	ACP1B*	0.6789	0.7067	0.6535	0.7790
	ACP1C*	0.0042	0.0032	0.0064	—

Values in parentheses are numbers tested.

Allele frequency data by Das et al. (1996).

Table 15.

Contribution of the Jat Sikh and Central Indian Tribal Populations to the Mazhabi and Ramdasi Sikh Scheduled Caste Populations of Punjab

	Contribution of parental population (%)
Hybrid population	Jat Sikh	Central Indian tribes
Mazhabi	67.42±11.90	32.58±11.90
Ramdasi	24.49±10.90	75.61±10.90

Discussion

The Indian region as a whole is characterized by relatively high frequencies of ABO*B, usually somewhat exceeding those of ABO*A (Mourant et al., 1976). This was true also for the Punjab, for in each of the 21 Sikh groups considered, incidence of ABO*B was consistently higher than that of ABO*A. Complete absence or rather low frequencies of RH*d characterize tribes of India inhabiting its central (range: nil-0.1392) (Das et al., 1996), southern (range: 0.1459-0.1803) (Papiha et al., 1997), and eastern (range: 0.0928-0.1615) (Papiha et al., 1988) states. Indeed, there was a reflection of these values in the Mazhabi and Ramdasi populations of Punjab, both of which showed similar low incidence of the allele (combined range: 0.1078-0.2331; Tables 6 and 7), suggesting their tribal antecedents.

In comparison to the Jat Sikh, both the Mazhabi and Ramdasi were generally characterized by higher frequencies of ESD*2, PGM1*2, AK1*2, and GLO1*1 and lower frequencies of ADA*2 and ACP1*A. The frequency of ESD*2 found in the Mazhabi and Ramdasi groups of Punjab was comparatively high (combined range: 0.1893-0.3315; Tables 6 and 7) and similar to the high value of the allele recorded in tribes of central (range: 0.2448-0.3989) (Das et al., 1996), southern (range: 0.1687-0.3120) (Papiha et al., 1997), and eastern (range: 0.313-0.336) (Papiha et al., 1988) India. Likewise, there was similarity in ACP1*A frequency between the present Sikh scheduled caste groups (combined range: 0.2317-0.3547; Tables 6 and 7) and tribes of central (range: 0.1573-0.2905), southern (range: 0.1862-0.4643), and eastern (range: 0.219-0.234) India, all characterized by comparatively low value of the allele. The distribution of ACP1*C in India is essentially nontribal (Roberts et al., 1980) and the allele was not reported in any of 16 tribes of Central India (Das et al., 1996), three tribes of southern India (Papiha et al., 1997), and two tribes of eastern India (Papiha et al., 1988). Absence of ACP1*C in half of the Sikh scheduled caste groups (Tables 6 and 7) suggested their tribal affinities.

Genetic distance analysis demonstrated, first, little or no genetic affinities between the Jat Sikh and both the Mazhabi and Ramdasi and, second, little genetic relationship between the latter two Sikh scheduled caste populations, providing genetic base to the social stratification prevailing in the Sikh community of Punjab. The apparent similarities in allele frequencies noted in both serological (RH(D)) and biochemical (ESD and ACP1) markers between lower strata Sikh populations of Punjab and tribes of central, southern, and eastern India suggested some genetic affinities between them, and genetic admixture analysis indicated tribal origin of Ramdasi. Officially, presently there is no scheduled tribe inhabiting Punjab, but like the anthropometric and morphological observations, respectively, of Risley (1915) and Jammu (1996) on Mazhabi and Ramdasi suggesting a higher element of indigenous extraction in them, this study provided some genetic evidence to support the theory of aboriginal tribal signatures in the Sikh scheduled caste population of the state.

Conclusions

Social stratification in the contemporary Sikh population of Punjab has a genetic basis, primarily attributable to different ethnic backgrounds of different castes. This lends support to Risley's century-old dictum that caste has a racial basis. There is evidence also for a higher element of autochthonous tribal extraction in the lower strata of Sikh population.

Footnotes

Acknowledgments

S.M.S.C. gratefully acknowledges University Grants Commission, New Delhi, for award of a Research Scientistship “B” in Science and for funds for the project entitled “Genetic Differentiation in Punjab Populations” (grant no. F 7-28(Sc.)/88(SA-I)).

Disclosure Statement

No competing financial interests exist.

References

Census of India (2001) Basic data sheet [table]

www.censusindia.gov.in/TablesPublished/Basic Data Sheet.aspx.

Chakraborty

. 1986. Gene admixture in human populations: models and predictions. Yrbk Phys Anthropol, 29:1-43.

Das

, Malhotra

, Mukherjee

et al. 1996. Population structure and genetic differentiation among 16 tribal populations of Central India. Hum Biol, 68:679-705.

Detter

, Ways

, Giblett

et al. 1968. Inherited variations in human phosphohexose isomerase. Ann Hum Genet, 31:329-338.

Ibbetson

. 1916. Punjab Castes. Government of India Press: Simla.

Jammu

. 1996. Society and culture in Punjab. Joshi

, Sidhu

, Kumar

et al. 83rd Indian Science Congress: Souvenir. Punjabi University: Patiala, India, 35-47.

Liu

. 2005. Power marker: an integrated analysis environment for genetic marker analysis. Bioinformatics, 21:2128-2129.

Mourant

. 1983. Blood Relations. Blood Groups and Anthropology. Oxford University Press: Oxford.

Mourant

, Kopec

, Domaniewska-Sobczak

. 1976. The Distribution of the Human Blood Groups and Other Polymorphisms, 2nd. Oxford University Press: London.

10.

Murch

, Gambel

, Kearney

. 1986. A double origin electrophoretic method for the simultaneous separation of adenosine deaminase, adenylate kinase and carbonic anhydrase II. J Forens Sci, 31:1349-1356.

11.

Nei

. 1972. Genetic distance between populations. Am Naturalist, 106:283-292.

12.

Nei

. 1973. Analysis of gene diversity in subdivided populations. Proc Nat Acad Sci USA, 70:3321-3323.

13.

Papiha

, Roberts

, Mishra

. 1988. Serogenetic studies among an urban and two tribal populations of Orissa, India. Ann Hum Biol, 15:143-152.

14.

Papiha

, Singh

, Lanchbury

et al. 1997. Genetic study of the tribal populations of Andhra Pradesh, South India. Hum Biol, 69:171-199.

15.

Risley

. 1915. People of India. Thacker, Spank & Co.: Kolkata.

16.

Roberts

, Papiha

, Rao

et al. 1980. A genetic study of some Andhra Pradesh populations. Ann Hum Biol, 7:199-212.

17.

Rose

. 1919. A Glossary of the Tribes and Castes of the Punjab and North-West Frontier Province, I. Imperial Book Depot: Delhi.

18.

Scott

, Fowler

JCS

. 1982. Electrophoretic typing of glyoxalase 1 (GLO1) isoenzymes using a mixed starch/agarose gel. Forens Sci Intern, 20:287-294.

19.

Sneath

PHA

, Sokal

. 1973. Numerical Taxonomy. Freeman: San Francisco.

20.

Thanseem

, Thangaraj

, Chaubey

et al. 2006. Genetic affinities among the lower castes and tribal groups of India: inference from Y chromosome and mitochondrial DNA. BMC Genetics, 7:42-53.

21.

Weir

, Cockerham

. 1984. Estimating F-statistics for the analysis of population structure. Evolution, 38:1358-1370.

22.

Wraxall

BGD

, Emes

. 1976. Erythrocyte acid phosphatase in blood stains. J Forens Sci Soc, 16:127-132.

23.

Wraxall

BGD

, Stolorow

. 1986. The simultaneous separation of the enzymes glyoxalase 1, esterase D and phosphoglucomutase. J Forens Sci, 31:1439-1449.

24.

Yasuda

. 1984. A note on gene frequency estimation in the ABO and ABO-like system. Jpn J Hum Genet, 29:371-380.