Abstract
Over one billion people worldwide suffer from obesity, and the number is continually rising. This epidemic is partly caused by the modern lifestyle. Animal models, especially mouse models, are crucial to identifying the genetic components of complex disorders and exploring the potential applications of these genetic findings. The body weight of the animals used in research is often measured regularly to monitor their health. Only endpoint measurements, such as ultimate body weight, are frequently examined in quantitative trait locus (QTL) studies; time series data, including weekly or biweekly body weight, are usually disregarded. QTL mapping using biweekly body weight measurements may be particularly intriguing in examining body weight gain in obesity research and identifying more genes associated with obesity and related metabolic disorders. This study is focused on identifying quantitative trait loci (QTLs) underlying body weight changes by analyzing biweekly weight measurements in collaborative cross (CC) mice maintained on a high‐fat diet for 12 weeks. QTL analysis, utilizing 525 mice from 55 CC lines (308 males and 217 females), revealed genome‐wide significant QTLs on different chromosomes for body weight changes over 12 weeks. This study unveiled 62 body weight QTLs, among which 28 novel QTLs associated with defined traits were observed and found not reported previously. In addition, 34 more QTLs were fine‐mapped, as the genomic interval positions of these had been previously identified. These findings highlight genomic regions that influence body weight in CC mice, underscoring the value of time series data in identifying novel genetic factors.
Keywords: candidate genes, collaborative cross mice, high‐fat diet, obesity, QTL mapping
Over one billion people worldwide suffer from obesity, and the number is continually rising. Animal models, especially mouse models, are crucial to identifying the genetic components of complex disorders and exploring the potential applications of these genetic findings. Only endpoint measurements, like ultimate body weight, are frequently examined in quantitative trait locus (QTL) studies; time series data, such as weekly or biweekly body weight, are usually disregarded. This study is focused on identifying quantitative trait loci (QTLs) underlying body weight changes by analyzing biweekly weight measurements in collaborative cross (CC) mice maintained on a high‐fat diet for 12 weeks. QTL analysis, utilizing 525 mice from 55 CC lines (308 males and 217 females), revealed genome‐wide significant QTLs on different chromosomes for body weight changes over 12 weeks. This study unveiled 62 body weight QTLs, among which 28 novels associated with defined traits were observed and found not reported previously. In addition, 34 more QTLs were fine‐mapped as the genomic interval positions of these were previously identified.
1. INTRODUCTION
The pathophysiology of obesity is known to be influenced by both inherited and environmental factors, and it differs significantly across populations and ethnic groups. Environmental influences, including those that may or may not contribute to obesity, can only be predicted by the host's genetic composition. Several genetic alterations induce a phenotypic alteration only under specific environmental conditions. Finding additional QTL for a more accurate analysis of obesity thus becomes a question in the QTL analysis of this condition. Consequently, to be included in this study, a search was conducted for unique body weight–specific QTL mapping in collaborative cross (CC) mice. Diseases related to gene–environmental interactions are commonly thought to be influenced by complex factors, including numerous host genes and obesity. According to current thinking, dietary decisions influence how genes are expressed, but genetics also plays a significant role in the development of one or more features of metabolic syndrome. This investigation focuses on the genes associated with the identified QTL in the subsequent stage.
More than one billion people in the world are obese, and the incidence is further increasing. 1 , 2 However, the stage is set by genetic variation, which influences the development and progress of the disease. Animal models, particularly mouse models, are crucial for elucidating the genetic contributions to complex diseases and for further investigating the function of these genetic discoveries. In most QTL investigations, only data such as final body weight as endpoint measurements are studied; time series data, such as weekly or biweekly body weight, are typically ignored. To investigate body weight gain in obesity research and identify additional genes associated with obesity and related metabolic disorders that exhibit their effect more prominently during development or disease progression and may fade or disappear later, QTL mapping using biweekly body weight measurements may be particularly interesting. The next‐generation mouse genetic reference population (GRP), known as the CC mice, is based on the use of various wild‐derived strains and outbred mice. It is believed that the wild‐derived strains contributed to the enrichment of genetic diversity during the generation of the CC lines. 3 , 4 , 5
Since the development of genome‐wide association studies (GWAS), the identification of obesity susceptibility variants has improved. Understanding the etiology of such complicated characteristic disorders as obesity and metabolic syndrome, therefore, requires identifying genetic markers for these conditions in populations with similar environmental exposure. In the current report, we fine‐mapped previously identified QTL associated with body weight increase. We went on to show how to use extensive phenotypic data on body weight development assessment to identify new QTL and genes associated with body weight change in response to a high‐fat diet (HFD). In the current study, we focused on biweekly visits in this CC population from week 12 to week 20. These data allowed us to identify body weight QTL that contributes to the overall obese phenotype peculiar to the CC lines and the known major QTL on different chromosomes.
2. MATERIALS AND METHODS
2.1. Ethics statement
According to the standard protocol authorized by the Animal Use and Care Committee at Tel Aviv University (TAU), mice were housed in the Faculty of Medicine's small animal facility (approved experiment number M‐12‐025).
2.2. CC lines, animal housing, and diet
The mice were housed in open‐top cages with hardwood chip bedding, separated by gender and CC line, and maintained under a 12 h light/dark cycle. Table S1 displays the specifics of the mice employed and the distribution of all the CC lines in each sex.
2.3. Genotyping and phenotyping
All CC lines were genotyped with a high‐density custom‐designed Mouse Universal Genotyping Array (MegaMUGA)—built on an Illumina platform consisting of approximately 230 000 single nucleotide polymorphisms (SNPs), which are available at http://mtweb.cs.ucl.ac.uk/mus/www/preCC/MEGA_MUGA/Mar2015.MEGA+MDA+MUGA/. 6
2.4. QTL mapping
Data analysis was performed using the statistical software R, including the HAPPY package. 6 Briefly, the HAPPY package also tests for the existence of a QTL at each locus using the estimated probabilities of descent from the founder strains to estimate the phenotypic effect attributable to each founder strain. If these effects are significantly different, then there is evidence for QTL. Thus, for the initial stage of linkage analysis based on single‐trait analysis, a standard polyallelic “marker” model will be used. This model tests to determine whether there is an overall association between the marker and the trait by comparing the results after fitting the complete model, which includes the mean, any fixed effects such as generation, and the marker alleles. Initially, we performed 1000 permutation tests to determine the statistical genome‐wide significance thresholds of −log10 (p)‐values corresponding to 95%, 90%, and 50% of the marker‐based heritability of the trait. Subsequently, the 95%, 90%, and 50% confidence intervals (CIs), which are above the threshold values, provided the genomic intervals in which the marker associated with the assessed trait is likely to be mapped.
2.5. Founder's effect analysis and statistical analysis
Sequence variations that segregate in the CC should segregate in the CC founders as well, except for a few de novo mutations that arise during breeding. We examined which variations matched the pattern of activity at the QTL under QTL peak. 6
2.6. Gene prioritization and functional analysis
The Mouse Genome Informatics (MGI) database resource, available at https://www.informatics.jax.org/, was utilized for the functional interpretation and gene and cellular pathway prioritization analysis of the QTL results. 7
3. RESULTS
After 12 weeks of the HFD challenge, the obesity effect was assessed to determine the response of the CC lines. Using the HAPPY software, the genotypic data of the CC lines were analyzed to determine any genetic correlation with the body weight features for both the male and female populations and the entire population.
3.1. Significant body weight variation between CC lines
Figure 1A–C illustrates the significant differences in percentage body weight means between the entire (overall) CC line population (Figure 1A) and for males and females, as shown in Figure 1B,C, respectively. The mean body weight change for the entire population of CC lines (n = 525/55 CC lines) was observed in varying degrees, with the highest mean values in IL2131, and the mean values from IL2668 representing the lowest value in the overall population. In the male‐only group, the mean body weight change ranged from the highest mean values in IL2131, with the mean values from IL2462 representing the lowest value. The mean body weight change ranged from the highest mean values in IL2146, with the mean values from IL2668 representing the lowest in the female‐only group.
FIGURE 1.
Total time series data of body weight at different time points of the mice age of different collaborative cross (CC) lines in response to high‐fat diet (HFD, 42%) challenge. Bar graph (A) shows the body weight of total time series data of 55 CC lines of the overall CC lines. Bar graph (B) shows the body weight of total time series data of females from 41 CC lines. Bar graph (C) shows the body weight of total time series data of males from 52 CC lines. The X‐axis represents the line number of different CC lines. The Y‐axis represents the total time series body weight data (min × mg/dL).
3.2. Identification of candidate genes within the mapped QTL intervals
A total of 62 QTLs were identified in our study and designated as ObSL (Obesity‐specific loci) and numbered. The mapped QTLs and their genomic interval locations, peak interval sizes, chromosome specificity, and specific traits are listed in Tables 1, 2, 3, and Figure S1A–C, respectively, representing the overall population, males, and females separately. A total of 28 QTLs were not previously mapped. The genomic intervals of the remaining 34 QTLs were previously reported, and their genomic intervals were fine‐mapped in our study, as shown in Table S2. The genomic positions of 42 of the mapped QTLs overlapped (clustered) on seven different peaks or close genomic positions, as shown in Table S3.
TABLE 1.
Summary of significant ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]) for the overall population.
| Trait | Chromosome | QTL | Peak (Mb) | CI 50% Size (Mb) (genes) | CI 90% Size (Mb) (genes) | CI 95% Size (Mb) (genes) |
|---|---|---|---|---|---|---|
| ΔBW0‐2 | chr14 | ObSL1* | 57.24 |
56.57–57.51 0.94 (32) |
53.94–59.34 5.40 (267) |
52.87–61.00 8.12 (414) |
| chr15 | ObSL2** | 28.45 |
27.83–29.88 2.05 (8) |
24.68–35.34 10.67 (104) |
23.64–37.32 13.68 (145) |
|
| chr15 | ObSL3* | 54.46 |
53.38–55.65 2.27 (31) |
46.50–59.75 13.25 (123) |
44.88–62.17 17.29 (158) |
|
| ΔBW0‐4 | chr11 | ObSL4** | 19.05 |
18.39–19.27 0.88 (15) |
16.81–20.29 3.48 (52) |
15.84–20.80 4.96 (72) |
| chr16 | ObSL5** | 49.36 |
47.87–50.47 2.60 (37) |
42.65–56.35 13.69 (176) |
40.98–58.17 17.19 (211) |
|
| chr16 | ObSL6** | 75.55 |
73.73–77.92 4.19 (39) |
66.54–84.62 18.09 (121) |
65.73–85.32 19.58 (140) |
|
| ΔBW0‐8 | chr5 | ObSL7** | 58.57 |
57.18–60.17 2.99 (14) |
53.02–65.27 12.25 (86) |
51.48–67.14 15.65 (147) |
| chr15 | ObSL8** | 28.45 |
28.17–29.05 0.87 (2) |
25.69–31.48 5.79 (51) |
24.83–32.46 7.63 (68) |
|
| ΔBW0‐10 | chr5 | ObSL9** | 50.85 |
49.38–52.80 3.42 (26) |
44.19–57.56 13.38 (98) |
42.17–59.08 16.91 (129) |
| chr5 | ObSL10** | 56.40 |
55.26–57.88 2.62 (15) |
51.82–62.10 10.28 (64) |
50.27–64.20 13.93 (88) |
|
| ΔBW0‐12 | chr1 | ObSL11* | 18.08 |
17.88–18.61 0.73 (8) |
15.77–20.78 5.01 (45) |
14.54–21.74 7.20 (76) |
| chr5 | ObSL12** | 53.33 |
50.67–55.07 4.39 (38) |
45.30–60.96 15.67 (103) |
44.11–62.51 18.40 (117) |
|
| chr11 | ObSL13** | 18.24 |
17.68–18.46 0.78 (7) |
15.58–19.44 3.86 (48) |
14.71–20.08 5.37 (61) |
|
| ΔBW2‐4 | chr2 | ObSL14* | 118.08 |
117.48–118.53 1.05 (17) |
113.41–121.34 7.94 (178) |
111.83–122.67 10.84 (273) |
| chr16 | ObSL15* | 49.61 |
45.39–51.30 5.91 (82) |
40.14–57.93 17.79 (213) |
39.69–59.10 19.41 (248) |
|
| chr16 | ObSL16** | 75.44 |
74.37–77.67 3.30 (32) |
68.53–83.95 15.42 (105) |
66.36–84.97 18.61 (132) |
|
| ΔBW6‐8 | chr1 | ObSL17** | 62.39 |
61.92–62.81 0.89 (7) |
60.14–64.32 4.18 (72) |
59.13–64.85 5.73 (106) |
| chr1 | ObSL18** | 32.96 |
31.00–33.84 2.84 (32) |
28.28–36.11 7.83 (86) |
26.82–37.24 10.42 (120) |
|
| chr12 | ObSL19** | 29.26 |
26.12–30.00 3.88 (32) |
19.97–37.75 17.78 (236) |
19.26–38.85 19.59 (255) |
|
| chr19 | ObSL20* | 54.50 |
52.47–55.21 2.75 (30) |
45.61–58.78 13.17 (170) |
44.80–60.49 15.68 (229) |
|
| ΔBW8‐10 | chr10 | ObSL21* | 124.66 |
124.34–124.99 0.65 (2) |
123.11–126.16 3.05 (14) |
122.25–126.67 4.41 (25) |
| ΔBW10‐12 | chr13 | ObSL22* | 113.98 |
112.60–115.71 3.11 (54) |
107.42–120.23 12.81 (193) |
105.81–122.14 16.33 (217) |
| ΔBW6‐12 | chr1 | ObSL23* | 58.57 |
58.16–59.24 1.08 (30) |
56.33–62.25 5.92 (108) |
53.66–63.28 9.62 (165) |
| chr1 | ObSL24** | 40.01 |
39.47–40.58 1.11 (24) |
35.90–43.34 7.45 (126) |
33.90–45.91 12.01 (198) |
|
| chr2 | ObSL25* | 57.50 |
56.75–58.59 1.85 (29) |
54.14–61.94 7.80 (95) |
52.76–64.45 11.69 (127) |
|
| chr5 | ObSL26** | 59.83 |
57.27–62.83 5.56 (24) |
51.48–69.04 17.56 (165) |
50.35–69.67 19.31 (177) |
|
| chr6 | ObSL27** | 94.57 |
93.03–98.54 5.51 (59) |
86.98–103.97 16.98 (243) |
85.56–104.52 18.97 (288) |
|
| chr12 | ObSL28* | 10.47 |
7.94–12.95 5.01 (56) |
2.18–19.13 16.95 (198) |
1.27–20.40 19.13 (218) |
Note: The levels of genome‐wide significance thresholds were **95% and *90%.
TABLE 2.
Summary of significant ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]) for males only.
| Interval | Chromosome | QTL | Peak (Mb) | 50% CI Size (Mb) (genes) | CI 90% Size (Mb) (genes) | CI 95% Size (Mb) (genes) |
|---|---|---|---|---|---|---|
| ΔBW0‐2 | chr3 | ObSL29* | 81.65 |
81.28–83.74 2.46 (26) |
79.16–87.90 8.73 (125) |
77.92–89.51 11.59 (246) |
| chr5 | ObSL30** | 23.91 |
21.91–25.59 3.67 (105) |
16.89–31.48 14.59 (288) |
15.27–32.98 17.70 (343) |
|
| chr5 | ObSL31** | 55.18 |
52.53–58.90 6.37 (46) |
47.30 64.62 17.32 (114) |
45.93–65.12 19.18 (126) |
|
| chr7 | ObSL32** | 47.05 |
46.85–47.16 0.31 (11) |
45.56–48.17 2.60 (106) |
44.67–50.48 5.81 (213) |
|
| chr7 | ObSL33* | 34.19 |
33.94–34.61 0.67 (27) |
32.63–35.98 3.34 (100) |
31.32–37.86 6.54 (142) |
|
| chr11 | ObSL34* | 18.26 |
17.71–18.51 0.80 (9) |
16.45–19.49 3.04 (45) |
15.90–20.02 4.12 (54) |
|
| chr15 | ObSL35** | 28.44 |
28.10–29.03 0.93 (3) |
25.33–31.74 6.40 (59) |
24.20–34.43 10.22 (91) |
|
| chr15 | ObSL36** | 58.94 |
58.27–59.36 1.09 (23) |
55.45–61.84 6.38 (83) |
53.13–62.98 9.85 (118) |
|
| ΔBW0‐4 | chr5 | ObSL37** | 25.20 |
23.24–26.67 3.43 (104) |
17.78–32.51 14.72 (308) |
16.26–33.80 17.53 (351) |
| chr5 | ObSL38** | 56.60 |
55.07–58.23 3.16 (16) |
49.46–63.49 14.03 (80) |
47.99–65.35 17.36 (127) |
|
| chr11 | ObSL39** | 18.84 |
18.44–19.04 0.59 (13) |
17.44–19.83 2.38 (30) |
16.92–20.24 3.32 (50) |
|
| chr15 | ObSL40** | 28.04 |
27.83–28.40 0.56 (6) |
26.24–30.64 4.40 (37) |
24.70–31.64 6.93 (63) |
|
| ΔBW0‐6 | chr15 | ObSL41** | 28.04 |
27.76–28.55 0.79 (6) |
25.92–30.88 4.96 (39) |
24.86–31.75 6.88 (63) |
| ΔBW0‐8 | chr5 | ObSL42** | 56.27 |
54.90–57.73 2.83 (15) |
50.20–62.20 12.00 (71) |
48.46–64.05 15.58 (94) |
| chr15 | ObSL43** | 28.04 |
27.83–28.39 0.55 (6) |
26.23–30.14 3.90 (32) |
25.38–30.97 5.58 (45) |
|
| ΔBW0‐12 | chr5 | ObSL44** | 53.20 |
51.08–55.46 4.37 (37) |
45.78–61.13 15.35 (92) |
44.01–62.36 18.35 (120) |
| ΔBW2‐4 | chr2 | ObSL45** | 109.23 |
106.98–110.45 3.47 (37) |
102.44–116.66 14.21 (271) |
100.99–117.72 16.73 (304) |
| chr2 | ObSL46** | 115.19 |
113.58–116.57 2.99 (35) |
106.13–121.65 15.52 (332) |
105.48–123.63 18.15 (393) |
|
| ΔBW6‐8 | chr1 | ObSL47** | 62.67 |
62.24–62.97 0.73 (8) |
60.90–64.07 3.17 (50) |
60.18–64.38 4.19 (71) |
| chr7 | ObSL48* | 58.64 |
57.26–60.95 3.69 (241) |
50.55–67.88 17.32 (404) |
49.49–68.41 18.91 (419) |
|
| chr12 | ObSL49** | 29.17 |
26.78–29.87 3.09 (24) |
20.84–36.57 15.73 (218) |
19.76–38.17 18.41 (240) |
|
| ΔBW8‐10 | chr5 | ObSL50** | 76.80 |
75.44–78.47 3.02 (46) |
69.82–83.92 14.10 (152) |
68.07–85.42 17.34 (168) |
| ΔBW10‐12 | chr4 | ObSL51** | 104.57 |
102.78–106.55 3.76 (51) |
97.03–113.02 15.98 (297) |
95.56–114.02 18.46 (322) |
| ΔBW6‐12 | chr5 | ObSL52** | 76.65 |
73.14–78.54 5.40 (88) |
67.44–85.33 17.89 (180) |
66.81–86.38 19.57 (201) |
Note: The levels of genome‐wide significance thresholds were **95% and *90%.
TABLE 3.
Summary of significant ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]) for females only.
| Interval | Chromosome | QTL | Peak (Mb) | 50% CI Size (Mb) (genes) | CI 90% Size (Mb) (genes) | CI 95% Size (Mb) (genes) |
|---|---|---|---|---|---|---|
| ΔBW0‐4 | chr3 | ObSL53** | 132.27 |
132.01–132.62 0.60 (3) |
131.01–133.50 2.48 (32) |
130.39–133.99 3.60 (43) |
| ΔBW0‐6 | chr3 | ObSL54** | 132.34 |
132.15–132.61 0.46 (2) |
131.39–133.26 1.87 (20) |
130.98–133.58 2.60 (33) |
| ΔBW0‐8 | chr3 | ObSL55* | 132.29 |
132.11–132.65 0.53 (3) |
131.19–133.55 2.35 (29) |
130.81–133.85 3.04 (39) |
| chr16 | ObSL56** | 49.24 |
48.53–50.13 1.60 (25) |
45.27–53.50 8.23 (95) |
44.01–56.03 12.02 (147) |
|
| ΔBW0‐10 | chr3 | ObSL57** | 132.34 |
132.14–132.60 0.45 (2) |
131.26–133.22 1.95 (25) |
130.98–133.58 2.59 (33) |
| ΔBW0‐12 | chr3 | ObSL58** | 132.29 |
131.91–132.63 0.72 (5) |
130.56–133.58 3.01 (41) |
129.67–134.08 4.40 (58) |
| ΔBW6‐8 | chr1 | ObSL59* | 32.95 |
32.42–33.76 1.33 (19) |
29.94–35.96 6.02 (72) |
28.86–37.32 8.45 (112) |
| chr13 | ObSL60** | 8.75 |
8.48–9.01 0.52 (11) |
7.19–9.91 2.71 (29) |
6.58–10.28 3.70 (34) |
|
| ΔBW6‐12 | chr1 | ObSL61** | 38.18 |
37.62–39.54 1.92 (40) |
34.09–43.66 9.57 (154) |
32.12–46.94 14.82 (230) |
| chr18 | ObSL62** | 90.42 |
88.98–90.59 1.60 (17) |
85.98–91.67 5.68 (31) |
84.36–92.34 7.95 (49) |
Note: The levels of genome‐wide significance thresholds were **95% and *90%.
3.3. Mapping new QTL
Our results showed that 28 new QTLs were not mapped in previous reports associated with our phenotypic data. The genomic interval positions of 16 of these QTLs overlapped with previously mapped QTLs (Tables S2 and S3), whereas 12 are unique. The first new QTL ObSL1 mapped during our study was on chromosome 14 at 52.87–61.00 Mbp, peaking at 57.24 Mbp in the overall population at BW02. The other observed new QTL of ΔBW02 in the overall population were ObSL2 and ObSL3, both located on chromosome 15, with interval positions of 23.64–37.32 and 44.88–62.17 Mbp, respectively, and their respective peak positions at 28.45 and 54.46 Mbp. In ΔBW04 phenotype in the overall population, three new QTLs, ObSL4, ObSL5, and ObSL6, were identified, the former on chromosome 11 and the later two on chromosome 16. In ΔBW08 phenotype in the overall population, a single new QTL ObSL8 was identified on chromosome 15 with an interval position of 24.83–32.46 Mbp with its respective peak positions at 28.45 Mbp. In ΔBW012 phenotype in the overall population, a single new QTL ObSL13 was identified on chromosome 11 with an interval position of 14.71–20.08 Mbp, reaching its peak position at 18.24 Mbp. In the overall population with phenotype ∆BW2‐4, three new QTLs, ObSL14, ObSL15, and ObSL16, were identified on two different chromosomes with their respective interval positions of 118.83–122.67, 39.69–59.10, and 66.36–84.97 Mbp. In the overall population with phenotypes of ∆BW810 and ∆BW1012, two new single QTLs, ObSL21 and ObSL22, were identified, respectively, on chromosomes 10 and 13 in each phenotype. In the overall population with phenotype ∆BW612, a single new QTL ObSL27 was identified on chromosome 6 with an interval position of 85.56–104.52 Mbp and a peak position at 94.57 Mbp.
The other major group of mapped QTLs was from the male population. Here, in males on chromosome 3, the identified QTL at ΔBW02 was ObSL29 on an interval position of 77.92–89.10 Mbp, with its peak position at 81.65 Mbp. In ΔBW04, two new QTLs, ObSL39 and ObSL40, were identified on chromosomes 11 and 15, respectively, with the respective interval positions of 16.92–20.24 and 24.70–31.64 Mbp. For ΔBW06 phenotype, a new QTL ObSL41, with the interval position of 24.86–31.75 Mbp, peaked at 28.04 Mbp was identified on chromosome 15. With the ΔBW08 phenotype, a novel QTL, ObSL43, with its respective interval position of 25.38–30.97 Mbp, was identified on chromosome 15. In males with phenotype ∆BW24, two new QTLs, ObSL45 and ObSL46, were identified on chromosome 2 with the respective interval positions of 100.99–117.72 Mbp, reaching their peaks at 109.23 and 115.19 Mbp. In males with ∆BW10–12 phenotype, a new QTL ObSL51 was identified on chromosome 4 with the interval position of 95.56–114.02 Mbp, reaching its peak at 104.57. In females with phenotypes of ΔBW08 and ∆BW68, two new QTLs, ObSL56 and ObSL60, were identified, respectively, on chromosomes 16 and 13, with the respective interval positions of 44.01–56.03 and 32.12–46.94 Mbp, reaching their peaks at 49.24 and 8.75 Mbp.
3.4. Fine‐mapping of previously identified QTL
The other 34 QTLs observed during our study were also new, but they had already been identified by earlier researchers, although our study fine‐mapped their interval position. In the overall population, 14 fine‐mapped QTLs were ObSL7, ObSL9, ObSL10, ObSL11, ObSL12, ObSL17, ObSL18, ObSL19, ObSL20, ObSL23, ObSL24, ObSL25, ObSL26, and ObSL28 on different chromosomes (Table 2). Similarly, in the males‐only group, 12 fine‐mapped QTL were observed on five different chromosomes throughout the genome (Table 2). In the female population, eight fine‐mapped QTLs were observed: ObSL53, ObSL54, ObSL55, ObSL57, ObSL58, ObSL59, ObSL61, and ObSL62, located on chromosomes 1, 3, and 13 (Table 2). Regarding the genome‐wide significance threshold, three sets of QTLs were identified during our study (Figure S1A–C). The first group had a 50% genome‐wide significance threshold. The second group had a 90% genome‐wide significance threshold, whereas the third group had a 95% genome‐wide significance threshold.
3.5. Founder effect
The results of this analysis are presented in Figure S2A–C, respectively, for the overall population, males only, and females only. We evaluated the effects of each founder genotype on the overall population, males, and females ΔBW mapped QTL interval, and estimated the assessed trait. The wild‐derived strains, primarily PWK, were a significant factor in the rise in ΔBW values, as demonstrated by the loci complicated pattern of founder haplotype effects, except in females where NOD plays the same role instead. Although other strains also contributed (positively or negatively) to the overall QTL effect, the NZO and NOD genotypes decreased this characteristic.
3.6. Gene prioritization and functional analysis
The results of the overall population are presented in Table 1 and Figure S1A.
For the DBW0‐2 phenotype, three QTLs (ObSL1, ObSL2, ObSL3), one on chromosome 14 and two on chromosome 15, containing 414, 145, and 158 genes in the 95% CI, respectively, were reported.
Interestingly, one of the systemic processes highly enriched in the MGI mammalian phenotype for this early time point is abnormal prenatal growth/weight/body size and from gene ontology lipid storage. Exostosin glycosyl transferase 1 (Ext1) is the highest prioritized gene. Ext1 is downregulated by a HFD, and this downregulation is considered to exacerbate endoplasmic reticulum (ER) stress and nonalcoholic fatty liver disease (NAFLD) progression. 8 Other prioritized genes include fat storage‐inducing transmembrane protein 1 (Fitm1) and hyaluronan synthase 2 (Has2). Has2 overexpression in mice has been shown to reduce fat accumulation and improve glucose tolerance. 9
For the DBW6‐12 phenotype, six QTLs (ObSL23, ObSL24, ObSL25, ObSL26, ObSL27, ObSL28), two on chromosome 1, one on chromosome 2, one on chromosome 5, one on chromosome 6, and one on chromosome 12, containing 165, 198, 127, 177, 288, and 218 genes in the 95% CI, respectively, were reported. Enriched systemic processes include increased body size and abnormal pituitary hormone level with notable genes prioritized being gonadotropin‐releasing hormone receptor (Gnrhr), ATPase, aminophospholipid transporter (APLT), class I, type 8A, member 1 (Atp8a1), phosphoribosyl pyrophosphate amidotransferase (Ppat), and neuromedin U (NmU). Nmu −/− mice are more susceptible to obesity, 10 and homozygous carriers of NMU gene mutations have increased the prevalence of overweight and obesity. 11
The results for the male population are presented in Table 2 and Figure S1B. For the DBW0‐2 phenotype, eight QTLs (ObSL29, ObSL30, ObSL31, ObSL32, ObSL33, ObSL34, ObSL35, ObSL36), one on chromosome 3, two on chromosome 5, two on chromosome 7, one on chromosome 11, and two on chromosome 15, containing 246, 343, 126, 126, 213, 142, 54, 91, and 118 genes in the 95% CI, respectively, were reported. The enriched systemic processes include impaired glucose tolerance and abnormal adipose tissue physiology and involved several highly prioritized genes shared with the overall population QTL, such as Ppargc1a.
The results of the female population are presented in Table 3 and Figure S1C. For female mice, we observed QTL on chromosome 3 for phenotypes DBW0‐4, DBW0‐6, DBW0‐8, DBW0‐10, and DBW0‐12 (ObSL53, ObSL54, ObSL55, ObSL57, ObSL58), showing strong involvement of the locus in many different time points. ObSL58 has been previously described by our laboratory as ObFL in Binenbaum et al. 12
4. DISCUSSION
To better comprehend the variations in body weight between the many CC lines exhibiting genetic diversity, 12 , 13 , 14 we looked into the cross‐population between several CC mice lines. These lines are not just closely linked genetically, but they also share known obesity loci on separate chromosomes, which accounts for the total variation in obesity across all these lines. 15 Using QTL mapping to analyze the time series body weight data, we could pinpoint distinct body weight QTL on various chromosomes that explains population variance across varying periods, as reported previously. 15 , 16
Based on our previous studies and reports, as presented here, we have demonstrated that male and female mice respond differently to the development of obesity when maintained on a HFD. It was speculated that different genes could be involved in this variation. Indeed, our QTL mapping results have shown that male and female mice have different QTLs associated with obesity development. We leveraged the large, well‐studied cohort, enabling us to perform combined analyses for both males and females, as well as separate analyses for male and female mice. Given the large number of mice used, it became evident to use a multipronged approach for this study.
In response to the environmental and nutritional challenge of a HFD (42%), this work employs a novel CC mouse model to investigate the intricate genetic architecture of obesity. Additional research is necessary to fully comprehend the biological functions and mechanisms of the obesity‐associated loci thus far in human GWAS surveys. 17 , 18 , 19 Several genes inside of our QTL intervals have not been associated with obesity in humans. As a result of finding new genes linked to the feature under investigation, our findings are novel. Finding phenotype–genotype connections has been successful for recent human GWAS. Numerous benefits, particularly for GWAS studies, make mouse models desirable. These benefits include the availability of genetic experimental tools, mouse genomic data resources, ease of breeding, and the capacity to provide highly controlled environmental conditions. 20 CC mouse populations that resemble models of various human groups were identified to analyze phenotypic–genotypic relationships and highlight interpopulation heterogeneity regarding similar environmental exposures.
After the QTL is mapped, basic and clinical researchers will use it extensively. Subsequent steps include candidate gene search and validation, translation of functional signaling pathways to high‐level human genetics, and candidate gene translation. The top candidate genes for the novel QTL on chromosome 15 associated with body weight are Gpt, Cbx6, Apol6, and Apol8. Gpt is an essential gene that codes for alanine aminotransferase 1 (ALT1), facilitating amino acid transamination into gluconeogenesis and the urea cycle. For the QTL on chromosome 16, Trap1 (TNF receptor‐associated protein 1) and Rrn3 (RNA polymerase I transcription factor homolog) were ranked as the top two candidate genes, which have been previously described in Delpero et al., 16 to be associated with final body weight at week 20.
Many factors, including environmental and genetic factors, influence body weight. Although some genes are known for their significant influence on body weight when dysfunctional (e.g., leptin, leptin receptor, MC4R, POMC), most genes have a small impact, and many more are still undiscovered. 2 Genes with minor effects are more complex to track, especially when their effect is time dependent, so they mainly act during specific periods, such as puberty. Although many environmental factors are well known, the contribution of many genetic factors and the interaction between genetic determinants and the environment are still unknown. 21 QTL mapping under different conditions, such as time series data, genetic background, and dietary condition, and the subsequent identification of genomic regions and candidate genes influencing obesity in mice and humans are essential to help understand the genetic contribution and its interaction with environmental factors in this common complex human disease. Finally, in our study, over 62 new QTLs were identified, which were thoroughly researched in the literature. Among these, 28 QTLs associated with defined traits were observed and found not to be reported previously, whereas an additional 34 QTLs were fine‐mapped.
It is often recognized that when it comes to obesity, genotypes have a significant impact on phenotypes. The findings of this study provide more evidence that phenotypes are influenced by genetics. Overall obesity is influenced by a number of hereditary factors, and the current study's findings are in line with earlier research in humans and mice. 22 Moreover, as highlighted in a recent review on leveraging mouse resources for investigating complex characteristics, available data in the eight founder strains investigated here support heritability of body weight. Our findings show how body weight and other particular parameters are distributed and heritable. The eight founder strains of the CC and diversity outbred mouse populations differed significantly for each parameter, and we also observed considerable heritability. After adjusting for body weight, most heritability estimates stayed quite high, indicating the role of genetics in obesity.
AUTHOR CONTRIBUTIONS
Hanifa J. Abu‐Toamih Atamni: Conceptualization; data curation; formal analysis. Iqbal M. Lone: Data curation; formal analysis; methodology; writing – original draft. Ilona Binenbaum: Data curation; formal analysis; methodology; software; writing – original draft. Kareem Midlej: Data curation; formal analysis. Eleftherios Pilalis: Data curation; formal analysis; supervision. Richard Mott: Conceptualization; data curation; formal analysis; funding acquisition; resources; writing – original draft. Aristotelis Chatziioannou: Data curation; formal analysis; software; validation; writing – original draft. Fuad A. Iraqi: Conceptualization; data curation; funding acquisition; investigation; methodology; project administration; resources; software; supervision; validation; visualization; writing – original draft; writing – review and editing.
FUNDING STATEMENT
This work was supported by the Hendrech and Eiran Gotwert Fund for studying diabetes, Wellcome Trust grants 085906/Z/08/Z, 075491/Z/04 and 090532/Z/09/Z, core funding by Tel‐Aviv University (TAU), Israeli Science Foundation (ISF) grants 429/09, 961/15 and 1085/18, United States‐Israel Binational Science Foundation (BSF) grant 2015077, German Israeli Science Foundation (GIF) grant I‐63‐410.20‐2017. Ilona Binenbaum's PhD thesis was supported by a scholarship from the State Scholarship Foundation in Greece (Athens) (Operational Program “Human Resources Development—Education and Lifelong Learning” Partnership Agreement (PA) 2014–2020).
CONFLICT OF INTEREST STATEMENT
Aristotelis Chatziioannou and Eleftherios Pilalis are cofounders of the e‐NIOS Applications PC. Besides, no other competing financial interests or associations might pose a conflict of interest (e.g., pharmaceutical stock ownership, consultancy). Fuad A. Iraqi is an editorial board member of Animal Models and Experimental Medicine (AMEM) and a corresponding author of this article. To minimize bias, he was excluded from all editorial decision making related to the acceptance of this article for publication.
Supporting information
Supplementary Figure 1. Genome scan of quantitative trait loci (QTLs) associated with total time series data of body weight at different time points of the mice age of different collaborative cross (CC) lines in response to high‐fat diet (HFD, 42%) challenge in the overall population (A), males only (B), and females only (C) in a population of 55 CC lines after 12 weeks on high‐fat (42% fat) dietary challenge. The X‐axis represents the 19 mouse chromosomes and the position of mapped QTL on the chromosome. The Y‐axis represents the logP of the test of association between locus and body weight trait. QTL based on permutation genome‐wide test, a significant level of p ≤ 0.05, was identified.
Supplementary Figure 2. The estimated effect size on the total time series data of body weight at different time points for the eight collaborative cross (CC) founder strains for different chromosome quantitative trait locus (QTL) in the overall population (A), male population only (B), and female population only (C). The X‐axis represents eight founder strains of the CC mice. The Y‐axis represents the haplotype effect size of the CC founder at the body weight QTL.
Supplementary Table 1. Summary of the used CC lines in our study. The name of each CC line is designated as IL#, which appears under the column CC lines. The number of males and females in each CC line is provided. Abbreviation: CC, collaborative cross.
Supplementary Table 2. Summary of significant Obesity‐Specific Locus, ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]) fine‐mapped by our study. The levels of genome‐wide significance thresholds were **95% and *90%.
Supplementary Table 3. Summary of significant Obesity‐Specific Locus, ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]), with overlapping of the genomic along with their linkage groups on each chromosome indicated by different colors. The levels of genome‐wide significance thresholds were **95% and *90%.
ACKNOWLEDGMENTS
None.
DATA AVAILABILITY STATEMENT
All CC lines were genotyped with a high‐density custom‐designed Mouse Universal Genotyping Array (MegaMUGA)—built on an Illumina platform consisting of approximately 230 000 SNPs, which are available at http://mtweb.cs.ucl.ac.uk/mus/www/preCC/MEGA_MUGA/Mar2015.MEGA+MDA+MUGA/. 6 The phenotype data are also available, upon request, from the corresponding author.
REFERENCES
- 1. Schwartz MW, Seeley RJ, Zeltser LM, et al. Obesity pathogenesis: an endocrine society scientific statement. Endocr Rev. 2017;38(4):267‐296. doi: 10.1210/er.2017-00111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Samblas M, Milagro FI, Martínez A. DNA methylation markers in obesity, metabolic syndrome, and weight loss. Epigenetics. 2019;14(5):421‐444. doi: 10.1080/15592294.2019.1595297 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Yehia R, Lone IM, Yehia I, Iraqi FA. Studying the pharmagenomic effect of Portulaca oleracea extract on anti‐diabetic therapy using the collaborative cross mice. Phytomed Plus. 2023;3(1):100394. doi: 10.1016/j.phyplu.2022.100394 [DOI] [Google Scholar]
- 4. Ghnaim A, Lone IM, Nun NB, Iraqi FA. Unraveling the host genetic background effect on internal organ weight influenced by obesity and diabetes using collaborative cross mice. Int J Mol Sci. 2023;24(9):1‐18. doi: 10.3390/ijms24098201 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Lone IM, Midlej K, Nun NB, Iraqi FA. Intestinal cancer development in response to oral infection with high‐fat diet‐induced type 2 diabetes (T2D) in collaborative cross mice under different host genetic background effects. Mamm Genome. 2023;34(1):56‐75. doi: 10.1007/s00335-023-09979-y [DOI] [PubMed] [Google Scholar]
- 6. Abu‐Toamih‐Atamni HJ, Lone IM, Binenbaum I, et al. Mapping novel QTL and fine mapping of previously identified QTL associated with glucose tolerance using the collaborative cross mice. Mamm Genome. 2024;2023(35):31‐55. doi: 10.1007/s00335-023-10025-0 [DOI] [PubMed] [Google Scholar]
- 7. Locke AE, Kahali B, Berndt SI, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518(7538):197‐206. doi: 10.1038/nature14177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Novelli G, Cassadonte C, Sbraccia P, Biancolella M. Genetics: a starting point for the prevention and the treatment of obesity. Nutrients. 2023;15(12):1‐16. doi: 10.3390/nu15122782 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cuciureanu M, Caratașu C‐C, Gabrielian L, et al. 360‐degree perspectives on obesity. Medicina (Kaunas). 2023;59(6):1‐51. doi: 10.3390/medicina59061119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Lone IM, Iraqi FA. Genetics of murine type 2 diabetes and comorbidities. Mamm Genome. 2022;33(3):421‐436. doi: 10.1007/s00335-022-09948-x [DOI] [PubMed] [Google Scholar]
- 11. Yosief RHS, Lone IM, Nachshon A. Identifying genetic susceptibility to Aspergillus fumigatus infection using collaborative cross mice and RNA‐seq approach. Animal Model Exp Med. 2024;7(1):36‐47. doi: 10.1002/ame2.12386 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Binenbaum I, Atamni HA‐T, Fotakis G, et al. Container‐aided integrative QTL and RNA‐seq analysis of collaborative cross mice supports distinct sex‐oriented molecular modes of response in obesity. BMC Genomics. 2020;21(1):761. doi: 10.1186/s12864-020-07173-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Lone IM, Nun NB, Ghnaim A, Schaefer AS, Houri‐Haddad Y, Iraqi FA. High‐fat diet and oral infection induced type 2 diabetes and obesity development under different genetic backgrounds. Animal Model Exp Med. 2023;6(2):131‐145. doi: 10.1002/ame2.12311 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lone IM, Zohud O, Midlej K, Brenner C. System genetic analysis of intestinal cancer and periodontitis development as influenced by aging and diabesity using collaborative cross mice. Animal Model Exp Med. 2025;8:758‐770. doi: 10.1002/ame2.12568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Baldarelli RM, Smith CL, Ringwald M, Richardson JE, Bult CJ, Mouse Genome Informatics Group . Mouse genome informatics: an integrated knowledgebase system for the laboratory mouse. Genetics. 2024;227(1):1‐31. doi: 10.1093/genetics/iyae031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Delpero M, Korkuć P, Arends D, Brockmann GA, Hesse D. Identification of additional body weight QTLs in the Berlin Fat Mouse BFMI861 lines using time series data. Sci Rep. 2024;14:6159. 10.1038/s41598-024-56097-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Krupke DM, Begley DA, Sundberg JP, Richardson JE, Neuhauser SB, Bult CJ. The mouse tumor biology database: a comprehensive resource for mouse models of human cancer. Cancer Res. 2017;77(21):e67‐e70. doi: 10.1158/0008-5472.CAN-17-0584 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Koutsandreas T, Binenbaum I, Pilalis E, Valavanis I, Papadodima O, Chatziioannou A. Analyzing and visualizing genomic complexity for the derivation of the emergent molecular networks. Int J Monit Surveill Technol Res. 2016;4(2):30‐49. doi: 10.4018/IJMSTR.2016040103 [DOI] [Google Scholar]
- 19. Wang M, Li B, Qin F, Ye J, Jin L. Obesity induced Ext1 reduction mediates the occurrence of NAFLD. Biochem Biophys Res Commun. 2022;589:123‐130. doi: 10.1016/j.bbrc.2021.12.017 [DOI] [PubMed] [Google Scholar]
- 20. Zhu Y, Li N, Huang M, et al. Adipose tissue hyaluronan production improves systemic glucose homeostasis and primes adipocytes for CL 316,243‐stimulated lipolysis. Nat Commun. 2021;12(1):4829. doi: 10.1038/s41467-021-25025-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Hanada R, Teranishi H, Pearson JT, et al. Neuromedin U has a novel anorexigenic effect independent of the leptin signaling pathway. Nat Med. 2004;10(10):1067‐1073. doi: 10.1038/nm1106 [DOI] [PubMed] [Google Scholar]
- 22. Hainerová I, Torekov SS, Ek J, et al. Association between neuromedin U gene variants and overweight and obesity. J Clin Endocrinol Metab. 2006;91(12):5057‐5063. doi: 10.1210/jc.2006-1442 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Genome scan of quantitative trait loci (QTLs) associated with total time series data of body weight at different time points of the mice age of different collaborative cross (CC) lines in response to high‐fat diet (HFD, 42%) challenge in the overall population (A), males only (B), and females only (C) in a population of 55 CC lines after 12 weeks on high‐fat (42% fat) dietary challenge. The X‐axis represents the 19 mouse chromosomes and the position of mapped QTL on the chromosome. The Y‐axis represents the logP of the test of association between locus and body weight trait. QTL based on permutation genome‐wide test, a significant level of p ≤ 0.05, was identified.
Supplementary Figure 2. The estimated effect size on the total time series data of body weight at different time points for the eight collaborative cross (CC) founder strains for different chromosome quantitative trait locus (QTL) in the overall population (A), male population only (B), and female population only (C). The X‐axis represents eight founder strains of the CC mice. The Y‐axis represents the haplotype effect size of the CC founder at the body weight QTL.
Supplementary Table 1. Summary of the used CC lines in our study. The name of each CC line is designated as IL#, which appears under the column CC lines. The number of males and females in each CC line is provided. Abbreviation: CC, collaborative cross.
Supplementary Table 2. Summary of significant Obesity‐Specific Locus, ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]) fine‐mapped by our study. The levels of genome‐wide significance thresholds were **95% and *90%.
Supplementary Table 3. Summary of significant Obesity‐Specific Locus, ObSL quantitative trait locus (QTL) for time series data of body weight traits at different time points of the mice age (following 12 weeks of high‐fat diet [HFD]), with overlapping of the genomic along with their linkage groups on each chromosome indicated by different colors. The levels of genome‐wide significance thresholds were **95% and *90%.
Data Availability Statement
All CC lines were genotyped with a high‐density custom‐designed Mouse Universal Genotyping Array (MegaMUGA)—built on an Illumina platform consisting of approximately 230 000 SNPs, which are available at http://mtweb.cs.ucl.ac.uk/mus/www/preCC/MEGA_MUGA/Mar2015.MEGA+MDA+MUGA/. 6 The phenotype data are also available, upon request, from the corresponding author.