The 16S rRNA gene amplicon sequencing is a popular technique that provides accurate characterization of microbial taxonomic abundances but does not provide any functional information. Several tools are available to predict functional profiles based on 16S rRNA gene sequence data that use different genome databases and approaches. As variable regions of partially-sequenced 16S rRNA gene cannot resolve taxonomy accurately beyond the genus level, these tools may give inflated results. Here, we developed ‘MicFunPred’, which uses a novel approach to derive imputed metagenomes based on a set of core genes only, thereby minimizing false-positive predictions. On simulated datasets, MicFunPred showed the lowest False Positive Rate (FPR) with mean Spearman's correlation of 0.89 (SD = 0.03), while on seven real datasets the mean correlation was 0.75 (SD = 0.08). MicFunPred was found to be faster with low computational requirements and performed better or comparable when compared with other tools.

16S rRNA 基因扩增子测序是一种流行的技术，可提供微生物分类丰度的准确表征，但不提供任何功能信息。有几种工具可用于基于 16S rRNA 基因序列数据预测功能图谱，这些数据使用不同的基因组数据库和方法。由于部分测序的 16S rRNA 基因的可变区域无法准确解析超出属水平的分类，这些工具可能会给出夸大的结果。在这里，我们开发了“MicFunPred”，它使用一种新颖的方法来推导仅基于一组核心基因的宏基因组，从而最大限度地减少假阳性预测。在模拟数据集上，MicFunPred 的误报率 (FPR) 最低，平均 Spearman 相关性为 0.89 (SD = 0.03)，而在七个真实数据集上，平均相关性为 0.75 (SD = 0.08)。发现 MicFunPred 速度更快，计算要求低，与其他工具相比表现更好或相当。