A Cancer Treatment Plan Recommendation System Based on Whole Genome Sequencing Data of Cancer Tissues *

With the continuous development of gene sequencing technology, new targets for cancer treatment are constantly being discovered. To help oncologists develop corresponding treatment plans based on the results of gene sequencing of cancer tissues, this study used Python to code a treatment plan prompting system. This program first aligns the sample sequence to the reference sequence using Bowtie 2 and BWA to generate a sequence alignment and map (SAM) file. The SAM file is converted to a binary file (BAM) and sorted using Samtools. A variant call file (VCF) is then generated using Bcftools from the sorted BAM. Then, the mutated genes are determined from the Catalogue Of Somatic Mutations In Cancer database (COSMIC v101, released 19-NOV-24). The information about clinical trials of targeted therapies for the gene mutation is retrieved from the database COSMIC Actionability V14. To test the program, we downloaded gene sequencing data of lung, colorectal, liver, gastric, and breast cancer from the NIH Sequence Read Archive (SRA) database, 4 cases of each kind of cancer. The following results were obtained by processing the data with the program: For lung cancer, the number of approved drug combinations with an Objective Response Rate (ORR_TREAT) of $70 \%$ or more was $19-20$, the related genes was 9-11, and the related mutation remark was 14-17. For colorectal cancer, the count ranges were 29-30, 16-17, and 2426, respectively. For liver cancer, the count ranges were 2125, 12-13, and 18-20, respectively. For gastric cancer, the count ranges were 29-30, 16-17, and 25-26. For breast cancer, the count ranges were $15-17,4-6$, and $8-12$. The above test results show that this program can provide clinicians with targeted treatment plans based on gene mutations in cancer tissues. Of course, the program is currently only a preliminary model and is still a long way from clinical application.