身為一個熱愛美食、喜歡在城市裡挖掘驚喜的人，臺中公益路一直是我最常出沒的地方之一。這條路可說是「臺中人的美食戰場」，從精緻西餐到創意火鍋，從日式丼飯到義式早午餐，每走幾步，就會有完全不同的特色料理餐廳。

這次我特別花了一整個月，實際造訪了公益路上十間口碑不錯的餐廳。有的是網友熱推的打卡名店，也有隱藏在巷弄裡的小驚喜。我以環境氛圍、口味表現、價格CP值與再訪意願為基準，整理出這篇實測評比。希望能幫正在猶豫去哪裡吃飯的你，找到那一間「吃完會想再來」的餐廳。

評比標準與整理方向

這次我走訪的10家餐廳橫跨不同料理類型，從高質感牛排館到巷弄系早午餐，每一間都有自己獨特的風格。為了讓整體比較更客觀，我依照以下四大面向進行評比，並搭配實際用餐體驗來打分。

整體而言，我希望這份評比不只是「哪家好吃」，而是幫你在不同情境下（約會、家庭聚餐、朋友小聚、商業午餐）都能快速找到合適的選擇。畢竟，美食不只是味覺的滿足，更是一段段與朋友共享的生活記憶。

10間臺中公益路餐廳評比懶人包

公益路向來是臺中人聚餐的首選地段，從火鍋、燒肉到中式料理與早午餐，每走幾步就有驚喜。以下是我實際造訪過的10間代表性餐廳清單，橫跨平價、創意、高級各路風格。

一頭牛日式燒肉｜炭香濃郁的和牛饗宴，約會聚餐首選

走在公益路上，很難不被 一頭牛日式燒肉 的木質外觀吸引。低調卻不失質感的門面，搭配昏黃燈光與暖色調的內裝，讓人一進門就感受到濃濃的日式職人氛圍。店內空間不大，但桌距規劃得宜，每桌皆設有獨立排煙設備，烤肉時完全不怕滿身油煙味。

餐點特色

一頭牛的靈魂，絕對是他們招牌的「三國和牛拼盤」。
嚴選的和牛部位，共八個部位、十樣餐點，讓人能從牛頭一路品嘗到牛尾。
油花分布均勻、切片厚薄恰好，經過炭火烤炙後香氣四溢，焦香與油脂在口中交融，入口即化的滑順感令人難忘。
值得一提的是，一頭牛的菜單設計十分彈性
想要一次體驗完整套餐也可以，偏好客製口味則能自由單點組合，不受套餐限制，想吃什麼就點什麼。
而且每桌都能選擇「自行燒烤」或「專人代烤」服務，代烤師的火侯掌握與節奏讓整體體驗更輕鬆愉快。
除了主角和牛，旬味野炊飯 與 主廚冰淇淋 也是隱藏版亮點，前者粒粒分明、香氣撲鼻；後者以香草與焙茶為基底，隨季節更換口味，完美收尾。整體服務親切熱情，特別是壽星還能享有 生日畫盤驚喜，讓慶祝時刻更添儀式感。

用餐體驗

整體節奏掌握得非常好。店員會在你剛想烤下一片肉時貼心遞上夾子、幫忙換烤網，讓人完全不用分心。整場用餐過程就像一場表演，從視覺、嗅覺到味覺都被滿足。
如果是第一次約會或慶祝特別節日，這裡的氛圍既不尷尬又不吵鬧，是營造氣氛的理想選擇。

綜合評分

地址：408臺中市南屯區公益路二段162號

電話：04-23206800

官網：http://www.marihuana.com.tw/yakiniku/index.html

小結語

一頭牛日式燒肉不僅是「吃肉的地方」，更像是一場五感盛宴。從進門那一刻到最後一道甜點，都能感受到他們對細節的用心。
若要在公益路找一間能讓人「邊吃邊微笑」的燒肉店，一頭牛 絕對值得列入你的必訪清單。

TANG Zhan 湯棧｜文青系火鍋代表，麻香湯底與視覺美感並重

在公益路這條美食戰線上，TANG Zhan 湯棧 是讓人一眼就會想走進去的那一種。
黑灰調的現代外觀、搭配微霧玻璃與招牌的「湯棧」燈字，呈現出一種低調的時尚感。
店內設計延續品牌主題，以「湯」為靈魂打造整體體驗，從裝潢到香氣，都有濃厚的溫潤氣息。

餐點特色

湯棧最有名的當然是它的「麻香鍋」。
湯底以雞骨與多種辛香料慢熬，香氣濃郁卻不嗆辣，入口後會在喉間留下柔和的花椒香。
「招牌麻油雞鍋」與「黃金牛奶鍋」也是人氣選項，特別是在冬天，溫潤的湯底配上滑嫩肉片，讓人每一口都覺得暖心。
他們的「滷肉飯」和「香蔥豆腐皮」更是許多老客人必點的靈魂配角，簡單卻有記憶點。

用餐體驗

整體氛圍比一般火鍋店更有質感。
桌距寬敞、燈光柔和，店員動作俐落又親切。即使客滿，也不會感覺吵雜或壓迫。
不論是一個人想靜靜吃鍋、或是朋友聚餐，湯棧都能給你剛剛好的距離與溫度。
值得一提的是，上菜速度快、湯底續湯毫不手軟，細節服務到位。

綜合評分

地址：408臺中市南屯區公益路二段248號

電話：04-22580617

官網：https://www.facebook.com/TangZhan.tw/

小結語

NINI 尼尼臺中店｜明亮寬敞的義式早午餐天堂

如果說前兩間是肉食愛好者的天堂，那 NINI 尼尼臺中店 絕對是想放鬆、聊聊天的好地方。餐廳外觀以白色系與大片玻璃窗為主，陽光灑進室內，讓人一踏入就有種度假般的輕盈感。假日早午餐時段特別熱鬧，建議提早訂位。

餐點特色

NINI 的菜單融合義式與臺灣人口味，選擇多樣且份量十足。主打的 松露燉飯 濃郁卻不膩口，米芯保留微Q口感；而 香蒜海鮮義大利麵 則以新鮮白蝦、花枝與淡菜搭配微辣蒜香，口感層次豐富。
此外，他們的薄餅披薩相當受歡迎，餅皮薄脆、餡料新鮮，是三五好友共享的好選擇。

用餐體驗

店內氣氛輕鬆不拘謹，無論是一個人帶電腦工作、或朋友聚餐，都能找到舒服角落。餐點上桌速度穩定，服務人員態度親切、補水與收盤都非常主動。整體節奏讓人覺得「時間變慢了」，很適合想遠離忙碌日常的人。

綜合評分

地址：40861臺中市南屯區公益路二段18號

電話：04-23288498

官網：https://nini.com.tw/

小結語

加分100%浜中特選昆布鍋物｜平價卻用心的湯頭系火鍋，家庭聚餐好選擇

在公益路這條高質感餐廳林立的戰場上，加分100%浜中特選昆布鍋物 走的是截然不同的路線。它沒有浮誇的裝潢、也沒有高價位的套餐，但靠著實在的湯頭與親切的服務，默默吸引許多回頭客。每到用餐時間，總能看到家庭或情侶三兩成群地圍著鍋邊聊天。

餐點特色

主打 北海道浜中昆布湯底，湯頭清澈卻不單薄，越煮越能喝出海藻與柴魚的自然香氣。
我這次點的是「牛奶昆布鍋」，入口時奶香與昆布香完美融合，搭配新鮮的牛五花肉片，滑順又不膩。
菜盤走健康取向，蔬菜比例高，連玉米、南瓜、豆皮都能吃出甜味；附餐的烏龍麵Q彈有嚼勁，吃完十分有飽足感。

用餐體驗

整體氛圍偏家庭取向，桌距寬敞、座位舒適，帶小孩來也不覺擁擠。店員態度親切，補湯、收盤都很勤快，給人一種「被照顧著」的安心感。
最難得的是，即使價位不高，食材新鮮度仍維持得很好，能感受到店家對品質的堅持。

綜合評分

地址：403臺中市西區公益路288號

電話：0910855180

官網：https://giafine100.com/

小結語

加分100%浜中特選昆布鍋物是一間「不浮誇、但會讓人想再訪」的火鍋店。它不追求豪華擺盤，而是用最簡單的湯頭與新鮮食材，傳遞出家常卻不平凡的溫度。
如果你想在公益路找一間可以放心帶家人一起吃的鍋物店，這裡絕對會讓人感到「加分」不少。

印月餐廳｜中式料理的藝術演繹，宴客與家庭聚會首選

說到臺中公益路的中式料理代表，印月餐廳 絕對是榜上有名。這間開業多年的餐廳以「中菜西吃」的概念聞名，把傳統中式料理以現代手法重新詮釋。從建築外觀到餐具擺設，每個細節都散發著低調的典雅氣息。
走進印月，挑高的空間、柔和的燈光與木質桌椅構成沉穩的氛圍。
不論是家庭聚餐、商務宴客，還是節日慶祝，都能找到恰到好處的格調。

餐點特色

印月最令人印象深刻的是他們將傳統中菜融入創意手法。
這次我品嚐的「松露雞湯」香氣濃郁、層次分明，一口下去既有中式的溫潤感，又帶出西式松露的奢華香氣。
「蒜香牛肋條」則是另一道招牌菜，外酥內嫩、油香十足，咬下去肉汁在口中散開，搭配特調醬汁非常過癮。
此外，他們的創意港點如「麻辣小籠包」與「金沙流沙包」也深受年輕客群喜愛，既保留經典又玩出新意。

用餐體驗

服務方面完全對得起餐廳的高級定位。從入座、點餐到上菜節奏，都拿捏得恰如其分。每道菜都會有服務人員細心介紹食材與吃法，讓人感受到「被款待」的尊榮感。
雖然價位偏中高，但在這樣的氛圍與品質下，物有所值。

綜合評分

地址：408臺中市南屯區公益路二段818號

電話：0422511155

官網：https://wein818.com/

小結語

印月餐廳是一間「不只吃飯，更像品味生活」的地方。
它成功地讓中式料理不再只是圓桌菜，而是能展現質感、講究細節的美食體驗。
若你在找一間能同時滿足味蕾與體面的餐廳，印月 絕對是公益路上的不敗經典。

KoDō 和牛燒肉｜極致職人精神，專為儀式感與頂級味覺而生

若要形容 KoDō 和牛燒肉 的用餐體驗，一句話足以總結——「像在欣賞一場關於肉的表演」。
隱身在公益路一隅，KoDō 的外觀低調典雅，店內以深色木質調與間接照明營造出沉穩氛圍。
從踏入店門那一刻開始，服務人員的態度、動線、聲音控制，全都精準到位，讓人彷彿走進日式劇場。

餐點特色

這裡主打 日本A5和牛冷藏肉，以「精切厚燒」的方式呈現。
我點的「壽喜燒風和牛套餐」是本日最驚艷的一道——服務人員現場以鐵鍋輕煎，再淋上特製壽喜燒醬汁，香氣瞬間瀰漫整桌。
肉片油花細緻、入口即化，搭配生蛋液後更添柔滑口感。
另一道「冷藏肋眼心」則保留了和牛的彈性與甜度，每一口都能感受到油脂與炭火交織出的層次。
即使是配角如「季節小菜」與「日式和風飯」也毫不馬虎，整體呈現出高級卻不造作的平衡。

用餐體驗

KoDō 的最大特色是「儀式感」。
每位店員的動作都有節奏，從擺盤、火候、換網到講解，都像排練過無數次的演出。
在這裡用餐，會自然地放慢速度，專注於每一口肉帶來的細膩變化。
特別推薦搭配店內的紅酒或日本威士忌，風味更加圓潤。

綜合評分

地址：403臺中市西區公益路260號

電話：0423220312

官網：https://www.facebook.com/kodo2018/

小結語

KoDō 和牛燒肉不是日常餐廳，而是一場體驗。
從環境、服務到食材，每個細節都讓人感受到對「完美」的執著。
若你想在公益路找一間能讓人留下深刻印象、適合紀念日慶祝的餐廳，KoDō 絕對是值得收藏的一次「味覺儀式」。

永心鳳茶｜在茶香裡用餐的優雅時光，臺味早午餐的新詮釋

走進 永心鳳茶公益店，彷彿進入一間有氣質的茶館。
柔和的燈光灑在復古綠牆上，搭配大理石桌面與金色餐具，整體氛圍既典雅又帶有一絲文青氣息。
這裡不只是喝茶的地方，更像是把「臺灣味」以早午餐的形式重新演繹。

餐點特色

永心鳳茶的餐點結合中式靈魂與西式擺盤，無論是「炸雞腿飯」還是「紅玉紅茶拿鐵」，都能讓人感受到熟悉卻不平凡的味道。
炸雞腿外酥內嫩，搭配自製酸菜與溏心蛋，鹹香中帶著層次感。
「鳳茶甜點拼盤」則以茶為靈魂——伯爵茶蛋糕、烏龍茶奶酪、紅茶雪酥，每一口都有細緻的香氣變化。
最特別的是他們的茶飲，從臺灣高山紅茶到金萱冷泡茶，每一壺都現泡現倒，香氣清雅。
對我而言，這不只是一頓飯，更是一段放鬆的午後儀式。

用餐體驗

店內服務人員態度溫和，對茶品介紹詳盡。上餐節奏剛好，不急不徐。
整體氛圍很「耐坐」，許多客人吃完正餐後仍會續點一壺茶聊天。
音樂輕柔、光線柔和，是那種可以靜靜待上兩小時的地方。

綜合評分

地址：40360臺中市西區公益路68號三樓(勤美誠品)

電話：0423221118

官網：https://linktr.ee/yonshin

小結語

永心鳳茶讓人重新定義「臺味」。
它不走傳統路線，而是把熟悉的元素以更細緻、更現代的方式呈現。
無論是姊妹下午茶、親子餐聚，或是想一個人沉澱片刻，永心鳳茶 都是一處能讓人慢下來、品味生活的好地方。

三希樓｜老饕級江浙功夫菜，穩重又帶人情味的中式饗宴

位於公益路上的 三希樓 是許多臺中老饕的口袋名單。
它沒有浮誇的裝潢，卻有一種低調的自信。從大門進入，就能聞到淡淡的醬香與蒸氣味，那是正宗江浙菜的靈魂。
整體裝潢以深木色為主，搭配圓桌與包廂設計，非常適合家庭聚餐或請客宴會。

餐點特色

三希樓的菜色以 江浙與港式料理 為主，兼顧傳統與現代風味。
我這次點了「東坡肉」與「蝦仁炒飯」，兩道都展現了主廚深厚的火候功力。
東坡肉油亮卻不膩，入口即化、鹹甜交織；蝦仁炒飯粒粒分明、香氣十足，每一口都吃得到鑊氣。
此外，「小籠包」皮薄多汁，是幾乎每桌必點的招牌；港點類如「金牌流沙包」與「干貝燒賣」也都表現穩定。

用餐體驗

三希樓的服務給人一種老派但貼心的感覺。
店員上菜節奏掌握得很好，會主動幫忙分菜、收盤，態度沉穩而不打擾。
最讓我印象深刻的是，這裡的客群非常多元——有帶長輩的家庭、公司聚餐，也有情侶共度節日，卻都能在同一空間裡感到自在。

綜合評分

地址：408臺中市南屯區公益路二段95號

電話：0423202322

官網：https://www.sanxilou.com.tw/

小結語

三希樓是一間「吃得出功夫」的餐廳。
它不追求創新，而是用穩定的味道與真材實料，抓住每一位饕客的胃。
如果你想在公益路上找一間能兼顧長輩口味、氣氛又不拘謹的中餐廳，三希樓 絕對是最穩妥的選擇。

一笈壽司｜低調奢華的無菜單日料，職人手藝詮釋旬味極致

在熱鬧的公益路上，一笈壽司 低調得幾乎不顯眼。
外觀簡約，沒有華麗招牌，只有小小的木質門面與柔黃燈光。
一推開門，迎面而來的是日式杉木香氣與寧靜的氛圍，吧檯座位整齊排列，主廚站在中間，彷彿舞臺上的演出者。

餐點特色

一笈壽司採 Omakase（無菜單料理） 形式，每一餐都由主廚根據當日食材設計。
我這次選擇中價位套餐（約 $1200），共十多道料理，從前菜、小鉢、刺身、握壽司到甜點一氣呵成。
「比目魚鰭邊握」是整場最驚豔的瞬間——主廚以火槍輕炙，油脂瞬間釋放，入口後化成柔滑香氣。
「甜蝦海膽軍艦」則完美展現鮮度與層次感，海膽甘甜、甜蝦緊實。
搭配主廚親自調配的醬汁，每一口都像在品嚐季節的節奏。

用餐體驗

整場用餐約90分鐘，節奏緩慢但沉穩。
主廚會邊料理邊與客人互動，介紹魚種產地與食材處理方式。
雖然整體空間不大，但氣氛極佳——柔和的音樂、清酒的香氣、刀刃切魚時的聲音，讓人完全沉浸其中。
特別喜歡他們最後的甜點「焙茶奶酪」，收尾清爽優雅，為整場體驗畫下完美句點。

綜合評分

地址：408臺中市南屯區公益路二段25號

電話：0423206368

官網：https://www.facebook.com/YIJI.sushi/

小結語

一笈壽司是一間真正讓人「放慢呼吸」的餐廳。
這裡沒有多餘擺盤，也不靠噱頭，而是以主廚對食材的尊重與技術堆疊出一場味覺饗宴。
若你想在公益路體驗日本料理最純粹的精神，一笈壽司 絕對值得你預約、靜靜期待。

茶六燒肉堂｜人氣爆棚的和牛燒肉聖地，肉香與幸福感同時滿分

若要票選公益路上「最難訂位」的餐廳，茶六燒肉堂 絕對名列前茅。
不管平日或假日，用餐時段幾乎一位難求。外觀以木質格柵搭配大面玻璃設計，呈現出年輕又有質感的風格。店內空間明亮、桌距適中，播放著輕快的音樂，整體氛圍熱鬧中帶點高級感，是許多年輕人聚餐、慶生的首選地。

餐點特色

茶六主打 和牛燒肉套餐，價格約落在 $700–$1000 間，份量與品質兼具。
我這次點的是「厚切牛舌套餐」，肉片厚實彈牙，略帶脆感，搭配鹽蔥提味剛剛好。
另一道「和牛拼盤」也相當受歡迎，油花分布均勻、香氣濃郁，輕烤幾秒即可入口即化。
套餐附餐部分也相當用心：沙拉新鮮、味噌湯濃郁，最後還有一份「茶香冰淇淋」作結尾，香氣清爽，完美收尾。

用餐體驗

茶六的服務效率相當高。店員親切、換網勤快、補水速度快，整場用餐流程流暢無壓力。
雖然客人很多，但環境維持得乾淨整潔，動線規劃良好。
最令人印象深刻的是他們的 整體節奏拿捏得剛剛好 ——餐點上桌快、氣氛熱絡，卻不會讓人覺得匆忙。
不論是朋友聚會、家庭聚餐，甚至是情侶約會，都能找到各自的樂趣。

綜合評分

地址：403臺中市西區公益路268號

電話：0423281167

官網：https://inline.app/booking/-L93VSXuz8o86ahWDRg0:inline-live-karuizawa/-LUYUEIOYwa7GCUpAFWA

小結語

茶六燒肉堂用「穩定品質＋輕奢氛圍」抓住了臺中年輕族群的心。
不論是第一次約會還是老朋友重聚，都能在這裡找到屬於燒肉的快樂節奏。
若你在公益路只想挑一家「保證不踩雷」的燒肉店，茶六燒肉堂 絕對是首選。

吃完10家公益路餐廳後的心得與結語

吃完這十家餐廳後，臺中公益路不只是一條美食街，而是一段生活風景線。

有的餐廳講究細膩與儀式感，像 一頭牛日式燒肉 與 一笈壽司，讓人感受到食材最純粹的美好

有的則以親切與溫度打動人心，像 加分昆布鍋物、永心鳳茶，讓人明白吃飯不只是為了飽足，而是一種被照顧的幸福。

而像茶六燒肉堂、TANG Zhan 湯棧 這類人氣名店，則用穩定的品質與熱絡的氛圍，成為許多臺中人心中「想吃肉就去那裡」的代名詞。

這十家店，構成了公益路最動人的縮影

有華麗的，也有溫柔的；有傳統的，也有創新的。

 每一家都在自己的風格裡發光，讓人吃到的不只是料理，而是一種生活的溫度與節奏。

對我而言，這不僅是一場美食旅程，更是一趟關於「臺中味道」的回憶之旅。

FAQ：關於臺中公益路美食常見問題

Q1：公益路哪一區的餐廳最集中？
 最熱鬧的區段大約在「公益路與黎明路口」一帶，這裡聚集了許多知名餐廳，從高級燒肉到早午餐通通有。
像 一頭牛日式燒肉、TANG Zhan 湯棧、茶六燒肉堂 都在這附近，交通方便、停車也相對容易。

Q2：需要提前訂位嗎？
 公益路的熱門餐廳幾乎都建議 提早3～5天訂位，尤其是假日或節慶期間。
特別是 一頭牛日式燒肉、KoDō 和牛燒肉、一笈壽司 這幾家，若臨時前往幾乎很難有位。

最後的話

若要用一句話形容這趟美食之旅，我會說：
「在公益路，吃飯不是選擇，而是一種享受。」
這條路上的每一次用餐，都像一段城市裡的小旅行。
下次當你不確定想吃什麼時，不妨沿著公益路走一圈，或許下一家，正好就是你新的最愛。

TANG Zhan 湯棧用餐環境舒服嗎？

如果你也和我一樣喜歡用味蕾探索一座城市，那就把這篇公益路美食攻略收藏起來吧。KoDō 和牛燒肉年末聚餐推薦嗎？

無論是約會、慶生、家庭聚餐，或只是想犒賞一下辛苦的自己——這條路上永遠會有一間剛剛好的餐廳在等你。印月餐廳春節期間適合來嗎？

下一餐，不妨從這10家開始。永心鳳茶尾牙預算好掌控嗎？

打開手機、約上朋友，讓公益路成為你生活裡最容易抵達的小確幸。茶六燒肉堂服務態度如何？

如果你有私心愛店，也歡迎留言分享，一頭牛日式燒肉公司聚餐適合嗎？

你的推薦，可能讓我下一趟美食旅程變得更精彩。茶六燒肉堂適合聚餐嗎？

This artistic rendering illustrates the diversity of mutational processes that generate clustered mutations in human cancer. Depicted here are kyklonas, which are molecular cyclones that cause mutations on circular extrachromosomal DNA (ecDNA), and omikli, which is a molecular fog that causes mutations on linear chromosomal DNA. Credit: Catherine Eng Mutation clusters play a role in 10% of human cancers. These clusters, caused by factors like APOBEC3 enzymes, accelerate cancer evolution and drug resistance on extrachromosomal DNA. Researchers led by bioengineers at the University of California San Diego have identified and characterized a previously unrecognized key player in cancer evolution: clusters of mutations occurring at certain regions of the genome. The researchers found that these mutation clusters contribute to the progression of about 10% of human cancers and can be used to predict patient survival. The findings were reported in a paper published on February 9, 2022, in the journal Nature. Clustered Somatic Mutations and Their Impact The work sheds light on a class of mutations called clustered somatic mutations—clustered meaning they group together at specific areas in a cell’s genome, and somatic meaning they are not inherited, but caused by internal and external factors such as aging or exposure to UV radiation, for example. Clustered somatic mutations have so far been an understudied area in cancer development. But researchers in the lab of Ludmil Alexandrov, a professor of bioengineering and cellular and molecular medicine at UC San Diego, saw something highly unusual about these mutations that warranted further study. “We typically see somatic mutations occurring randomly across the genome. But when we looked closer at some of these mutations, we saw that they were occurring in these hotspots. It’s like throwing balls on the floor and then suddenly seeing them cluster in a single space,” said Alexandrov. “So we couldn’t help but wonder: What is happening here? Why are there hotspots? Are they clinically relevant? Do they tell us something about how cancer has developed?” “Clustered mutations have largely been ignored because they only make up a very small percentage of all mutations,” said Erik Bergstrom, a bioengineering PhD student in Alexandrov’s lab and the first author of the study. “But by diving deeper, we found that they play an important role in the etiology of human cancer.” Mapping Cancer Mutations The team’s discoveries were enabled by creating the most comprehensive and detailed map of known clustered somatic mutations. They started by mapping all the mutations (clustered and non-clustered) across the genomes of more than 2500 cancer patients—an effort that in total encompassed 30 different cancer types. The researchers created their map using next-generation artificial intelligence approaches developed in the Alexandrov lab. The team used these algorithms to detect clustered mutations within individual patients and elucidate the underlying mutational processes that give rise to such events. This led to their finding that clustered somatic mutations contribute to cancer evolution in approximately 10% of human cancers. Taking it a step further, the researchers also found that some of the cancer-driving clusters—specifically those found in known cancer driver genes—can be used to predict the overall survival of a patient. For example, the presence of clustered mutations in the BRAF gene—the most widely observed driver gene in melanoma—results in better overall patient survival compared to individuals with non-clustered mutations. Meanwhile, the presence of clustered mutations in the EGFR gene—the most widely observed driver gene in lung cancer—results in decreased patient survival. “What’s interesting is that we see differential survival in terms of just having clustered mutations detected within these genes, and this is detectable with existing platforms that are commonly used in the clinic. So this acts as a very simple and precise biomarker for patient survival,” said Bergstrom. “This elegant work emphasizes the importance of developing AI approaches to elucidate tumor biology, and for biomarker discovery and rapid development using standard platforms with direct line of sight translation to the clinic,” said Scott Lippman, director of Moores Cancer Center and associate vice chancellor for cancer research and care at UC San Diego. “This highlights UC San Diego’s strength in combining engineering approaches in artificial intelligence for solving current problems in cancer medicine.” A New Mode of Cancer Evolution In this study, the researchers also identified various factors that cause clustered somatic mutations. These factors include UV radiation, alcohol consumption, tobacco smoking, and most notably, the activity of a set of antiviral enzymes called APOBEC3. APOBEC3 enzymes are typically found inside cells as part of their internal immune response. Their main job is to chop up any viruses that enter the cell. But in cancer cells, the researchers think that the APOBEC3 enzymes may be doing more harm than good. The researchers found that cancer cells—which are often rife with circular rings of extrachromosomal DNA (ecDNA) that harbor known cancer driver genes—have clusters of mutations occurring across individual ecDNA molecules. The researchers attribute these mutations to the activity of APOBEC3 enzymes. They hypothesize that APOBEC3 enzymes are mistaking the circular rings of ecDNA as foreign viruses and attempt to restrict and chop them up. In doing so, the APOBEC3 enzymes cause clusters of mutations to form within individual ecDNA molecules. This in turn plays a key role in accelerating cancer evolution and likely leads to drug resistance. The researchers named these rings of clustered mutations kyklonas, which is the Greek word for cyclones. “This is a completely novel mode of oncogenesis,” said Alexandrov. Along with the team’s other findings, he explained, “this lays the foundation for new therapeutic approaches, where clinicians can consider restricting the activity of APOBEC3 enzymes and/or targeting extrachromosomal DNA for cancer treatment.” Reference: “Mapping clustered mutations in cancer reveals APOBEC3 mutagenesis of ecDNA” by Erik N. Bergstrom, Jens Luebeck, Mia Petljak, Azhar Khandekar, Mark Barnes, Tongwu Zhang, Christopher D. Steele, Nischalan Pillay, Maria Teresa Landi, Vineet Bafna, Paul S. Mischel, Reuben S. Harris and Ludmil B. Alexandrov, 9 February 2022, Nature. DOI: 10.1038/s41586-022-04398-6 This work was supported by a Cancer Grand Challenge award from Cancer Research UK as well as funding from the U.S. National Institutes of Health, Alfred P. Sloan Foundation, and Packard Foundation.

Researchers have published the first evidence demonstrating the collection of animal environmental DNA (eDNA) from the air. Researchers from Queen Mary University of London have shown for the first time that animal DNA shed within the environment can be collected from the air. The proof-of-concept study, published in the journal PeerJ, opens up the potential for new ecological, health, and forensic applications of environmental DNA (eDNA), which to date has mainly been used to survey aquatic environments. Living organisms such as plants and animals shed DNA into their surrounding environments as they interact with them. In recent years, eDNA has become an important tool to help scientists identify species found within different environments. However, whilst a range of environmental samples, including soil and air, have been proposed as sources of eDNA until now most studies have focused on the collection of eDNA from water. In this study, the researchers explored whether eDNA could be collected from air samples and used to identify animal species. They first took air samples from a room that had housed naked mole rats, a social rodent species that live in underground colonies, and then used existing techniques to check for DNA sequences within the sampled air. Using this approach, the research team showed that airDNA sampling could successfully detect mole-rat DNA within the animal’s housing and from the room itself. The scientists also found human DNA in the air samples suggesting a potential use of this sampling technique for forensic applications. Opening Doors to Conservation and Forensics Dr. Elizabeth Clare, Senior Lecturer at Queen Mary University of London and first author of the study, said: “The use of eDNA has become a topic of increasing interest within the scientific community particularly for ecologists or conservationists looking for efficient and non-invasive ways to monitor biological environments. Here we provide the first published evidence to show that animal eDNA can be collected from air, opening up further opportunities for investigating animal communities in hard-to-reach environments such as caves and burrows.” The research team is now working with partners in industry and the third sector, including the company NatureMetrics, to bring some of the potential applications of this technology to life. Dr. Clare added: “What started off as an attempt to see if this approach could be used for ecological assessments has now become much more, with potential applications in forensics, anthropology, and even medicine.” “For example, this technique could help us to better understand the transmission of airborne diseases such as COVID-19. At the moment social distancing guidelines are based on physics and estimates of how far away virus particles can move, but with this technique, we could actually sample the air and collect real-world evidence to support such guidelines.” Reference: “eDNAir: proof of concept that animal DNA can be collected from air sampling” by Elizabeth L Clare, Chloe Economou, Chris G Faulkes, James D Gilbert, Frances Bennett, Rosie Drinkwater and Joanne E Littlefair, 31 March 2021, PeerJ. DOI: 10.7717/peerj.11030 The project was supported by Queen Mary’s Impact Acceleration Accounts (IAAs), strategic awards provided to institutions by UK Research and Innovation (UKRI) that support knowledge exchange (KE) and help researchers generate impact from their research.

Researchers at MIT and Tufts University have developed a new AI model called ConPLex that vastly accelerates drug discovery by predicting drug-protein interactions without the need to calculate the molecules’ structures. The model can screen over 100 million compounds in a single day, which could significantly reduce drug development failure rates and costs. By applying a language model to protein-drug interactions, researchers can quickly screen large libraries of potential drug compounds. Huge libraries of drug compounds may hold potential treatments for a variety of diseases, such as cancer or heart disease. Ideally, scientists would like to experimentally test each of these compounds against all possible targets, but doing that kind of screen is prohibitively time-consuming. In recent years, researchers have begun using computational methods to screen those libraries in hopes of speeding up drug discovery. However, many of those methods also take a long time, as most of them calculate each target protein’s three-dimensional structure from its amino-acid sequence, then use those structures to predict which drug molecules it will interact with. Researchers at MIT and Tufts University have now devised an alternative computational approach based on a type of artificial intelligence algorithm known as a large language model. These models — one well-known example is ChatGPT — can analyze huge amounts of text and figure out which words (or, in this case, amino acids) are most likely to appear together. The new model, known as ConPLex, can match target proteins with potential drug molecules without having to perform the computationally intensive step of calculating the molecules’ structures. Using this method, the researchers can screen more than 100 million compounds in a single day — much more than any existing model. “This work addresses the need for efficient and accurate in silico screening of potential drug candidates, and the scalability of the model enables large-scale screens for assessing off-target effects, drug repurposing, and determining the impact of mutations on drug binding,” says Bonnie Berger, the Simons Professor of Mathematics, head of the Computation and Biology group in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), and one of the senior authors of the new study. Lenore Cowen, a professor of computer science at Tufts University, is also a senior author of the paper, which was published on June 8 in the Proceedings of the National Academy of Sciences. Rohit Singh, a CSAIL research scientist, and Samuel Sledzieski, an MIT graduate student, are the lead authors of the paper, and Bryan Bryson, an associate professor of biological engineering at MIT and a member of the Ragon Institute of MGH, MIT, and Harvard, is also an author. In addition to the paper, the researchers have made their model available online for other scientists to use. Making Predictions In recent years, computational scientists have made great advances in developing models that can predict the structures of proteins based on their amino-acid sequences. However, using these models to predict how a large library of potential drugs might interact with a cancerous protein, for example, has proven challenging, mainly because calculating the three-dimensional structures of the proteins requires a great deal of time and computing power. An additional obstacle is that these kinds of models don’t have a good track record for eliminating compounds known as decoys, which are very similar to a successful drug but don’t actually interact well with the target. “One of the longstanding challenges in the field has been that these methods are fragile, in the sense that if I gave the model a drug or a small molecule that looked almost like the true thing, but it was slightly different in some subtle way, the model might still predict that they will interact, even though it should not,” Singh says. Researchers have designed models that can overcome this kind of fragility, but they are usually tailored to just one class of drug molecules, and they aren’t well-suited to large-scale screens because the computations take too long. The MIT team decided to take an alternative approach, based on a protein model they first developed in 2019. Working with a database of more than 20,000 proteins, the language model encodes this information into meaningful numerical representations of each amino-acid sequence that capture associations between sequence and structure. “With these language models, even proteins that have very different sequences but potentially have similar structures or similar functions can be represented in a similar way in this language space, and we’re able to take advantage of that to make our predictions,” Sledzieski says. In their new study, the researchers applied the protein model to the task of figuring out which protein sequences will interact with specific drug molecules, both of which have numerical representations that are transformed into a common, shared space by a neural network. They trained the network on known protein-drug interactions, which allowed it to learn to associate specific features of the proteins with drug-binding ability, without having to calculate the 3D structure of any of the molecules. “With this high-quality numerical representation, the model can short-circuit the atomic representation entirely, and from these numbers predict whether or not this drug will bind,” Singh says. “The advantage of this is that you avoid the need to go through an atomic representation, but the numbers still have all of the information that you need.” Another advantage of this approach is that it takes into account the flexibility of protein structures, which can be “wiggly” and take on slightly different shapes when interacting with a drug molecule. High Affinity To make their model less likely to be fooled by decoy drug molecules, the researchers also incorporated a training stage based on the concept of contrastive learning. Under this approach, the researchers give the model examples of “real” drugs and imposters and teach it to distinguish between them. The researchers then tested their model by screening a library of about 4,700 candidate drug molecules for their ability to bind to a set of 51 enzymes known as protein kinases. From the top hits, the researchers chose 19 drug-protein pairs to test experimentally. The experiments revealed that of the 19 hits, 12 had strong binding affinity (in the nanomolar range), whereas nearly all of the many other possible drug-protein pairs would have no affinity. Four of these pairs bound with extremely high, sub-nanomolar affinity (so strong that a tiny drug concentration, on the order of parts per billion, will inhibit the protein). While the researchers focused mainly on screening small-molecule drugs in this study, they are now working on applying this approach to other types of drugs, such as therapeutic antibodies. This kind of modeling could also prove useful for running toxicity screens of potential drug compounds, to make sure they don’t have any unwanted side effects before testing them in animal models. “Part of the reason why drug discovery is so expensive is because it has high failure rates. If we can reduce those failure rates by saying upfront that this drug is not likely to work out, that could go a long way in lowering the cost of drug discovery,” Singh says. This new approach “represents a significant breakthrough in drug-target interaction prediction and opens up additional opportunities for future research to further enhance its capabilities,” says Eytan Ruppin, chief of the Cancer Data Science Laboratory at the National Cancer Institute, who was not involved in the study. “For example, incorporating structural information into the latent space or exploring molecular generation methods for generating decoys could further improve predictions.” Reference: “Contrastive learning in protein language space predicts interactions between drugs and protein targets” by Rohit Singh, Samuel Sledzieski, Bryan Bryson, Lenore Cowen and Bonnie Berger, 8 June 2023, Proceedings of the National Academy of Sciences. DOI: 10.1073/pnas.2220778120 The research was funded by the National Institutes of Health, the National Science Foundation, and the Phillip and Susan Ragon Foundation.

RRG455KLJIEVEWWF