Average run time and memory usage to align 100 query samples and to reconstruct ancestral sequence for the SARS2-100k dataset.

We estimated the run time and peak memory usage for the two preprocessing steps by TIPars. The first step of query sample alignment is necessary for all other alignment-based phylogenetic placement methods, e.g., UShER, EPA-ng, IQTREE-2 and APPLES-2. For the second step, once ancestral sequences have been reconstructed, they can be reused. Here we performed 10 runs of these two preprocessing steps: 1) aligning 100 SARS-CoV-2 samples to the reference hCoV-19/Wuhan/WIV04/2019|EPI_ISL_402124 using MAFFT (—addtotop); 2) inferring the ancestral sequences for the SARS2-100k dataset using PastML (ACCTRAN method) and present average times. Tests were run on a server with 64 or 8 Intel Xeon Gold 6242 CPU cores.
(XLSX)

本研究通过TIPars评估了两项预处理步骤的运行时长与峰值内存占用。查询样本比对（query sample alignment）作为第一步，是所有其他基于比对的系统发育放置（phylogenetic placement）方法（如UShER、EPA-ng、IQTREE-2与APPLES-2）的必要前置步骤。至于第二步，祖先序列经重构后即可重复复用。本次实验共对两项预处理步骤执行10轮完整运行：1）使用MAFFT（参数--addtotop）将100株严重急性呼吸综合征冠状病毒2（SARS-CoV-2）样本比对至参考序列hCoV-19/Wuhan/WIV04/2019|EPI_ISL_402124；2）使用PastML的ACCTRAN法为SARS2-100k数据集推断祖先序列，并给出平均运行时长。所有测试均在搭载64核或8核Intel Xeon Gold 6242 CPU的服务器上完成。
(XLSX)