https://academic.oup.com/mbe/article/38/10/4149/5955855

Several neighboring populations within the Sino-Tibetan language group also displayed similar patterns of ancestry with the Taiwanese Han including Singapore Chinese (SG-CH), Chinese Cantonese (CN-GA), Chinese Han, and Tujia people (fig. 1B).

https://www.semanticscholar.org/paper/Mitochondrial-genetic-diversity-of-Pingpu-tribes-in-Shinoda-Kakuda/d64bb0b06d9fb887a5b26bc0324beb349f493f31

PC analysis reveals that the mtDNA haplogroup profiles of the Pingpu population are similar to the profiles of the populations from mainland southern China [Guangxi, Guangdong, Hainan] and North Vietnam and distinct from [Highland] aboriginal and Han Taiwanese. This is in agreement with the hypothesis that the Pingpu population has descended from immigrants from southern China, particularly Guangxi, and has a common origin with the populations from this region. It is said that the influence of the Pingpu by the Han Chinese immigrants began in the 17th century. However, our analysis indicates that their maternal genetic components continue to retain their original form.

https://academic.oup.com/mbe/article/38/10/4149/5955855

Our results are consistent with the findings of Chen et al. (2016), who utilized a smaller number of populations to identify four major ancestries and suggested that 80% of Taiwanese Han people are genetically closer to the Southern Han Chinese than to the Northern Han Chinese. However, the geographic patterns of these ancestries were not thoroughly discussed in their analysis. Although our inferred pattern of ancestries is also in good agreement with the previous studies that analyzed the Pan-Asia and HGDP data sets separately (Li et al. 2008; Abdulla et al. 2009), by analyzing the combined data, we were able to gain a better overview of the geographic distributions of these ancestries; consequently, they can be referred to as the Southeastern (blue), Northern (yellow), Island Southeast Asian (ISEA; pink), and Japonic (green) ancestries. Notably, we identified considerable proportions of ISEA ancestry (also carried by many Austronesian-speaking populations in high proportions) in most individuals of Taiwanese Han (average 15%, range 0.1 - 62%).

The mixed ancestries observed in the Taiwanese Han could be attributed to either population mixture or shared ancestry before the divergence of descendent populations. We therefore applied the F3 tests to detect signatures of population mixture. Consequently, our results showed that the ISEA ancestry in the Taiwanese Han was the outcome of population mixtures rather than shared ancestry, and the admixture event likely occurred before the Taiwanese Han ancestors migrated to Taiwan (fig. 2A). If the admixture occurred only after the Han people migrated to Taiwan, then the observed results would only be seen in the Taiwanese Han.

However, similar F3 outcomes were found in the Chinese Han (supplementary fig. S2, Supplementary Material online), supporting that admixture occurred prior to migration to Taiwan. Moreover, signatures of population admixture were also detected between the ancestors of Taiwanese Han and the Ami Austronesian-speaking population using the F4 test; significant positive F4 values were observed when most Sino-Tibetan speaking populations were individually included in the analysis, except for the Chinese Singapore and Chinese Cantonese (table 1). These two populations appear to be genetically closest to the Taiwanese Han among all other Sino-Tibetan speaking populations (fig. 1B), which is consistent with the hypothesis of population mixture before the ancestors of Taiwanese Han migrated to Taiwan.